diff --git a/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/2adbb622-d74e-47de-a015-51dcd2f952c3.json b/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/2adbb622-d74e-47de-a015-51dcd2f952c3.json new file mode 100644 index 0000000000000000000000000000000000000000..da108d182f892a744bae97da54aa3028f1ffd898 --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/2adbb622-d74e-47de-a015-51dcd2f952c3.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "56.0" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1669, + "details": { + "accuracy_raw": "16.69%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "6", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "56.0" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/3e193621-a254-44e2-8505-f3b38597010c.json b/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/3e193621-a254-44e2-8505-f3b38597010c.json new file mode 100644 index 0000000000000000000000000000000000000000..90946ac022ba62b5e6a7be9b8b39a3a4fcb3d3c9 --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-3-7-sonnet-20250219/3e193621-a254-44e2-8505-f3b38597010c.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "high", + "total_cost_usd": "16.13" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1308, + "details": { + "accuracy_raw": "13.08%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "8", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "16.13" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/5e6b50dd-3765-4fcc-9a94-3db4ee0f6787.json b/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/5e6b50dd-3765-4fcc-9a94-3db4ee0f6787.json new file mode 100644 index 0000000000000000000000000000000000000000..e05b4db1cdbbd37830b9a3b346c7b9a846ae8150 --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/5e6b50dd-3765-4fcc-9a94-3db4ee0f6787.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-opus-4-1-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 (August 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "385.43" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0726, + "details": { + "accuracy_raw": "7.26%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "11", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "385.43" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/f369d428-adf7-4a46-ba43-98dc4d227260.json b/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/f369d428-adf7-4a46-ba43-98dc4d227260.json new file mode 100644 index 0000000000000000000000000000000000000000..7714a994bdb0d77f33edde7b8f6cfe1716c87b07 --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-opus-4-1/f369d428-adf7-4a46-ba43-98dc4d227260.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-opus-4-1-high-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 High (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 High (August 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "high", + "total_cost_usd": "779.72" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1375, + "details": { + "accuracy_raw": "13.75%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "7", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "779.72" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/2825bc34-e49a-4342-8590-e899d26db9b6.json b/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/2825bc34-e49a-4342-8590-e899d26db9b6.json new file mode 100644 index 0000000000000000000000000000000000000000..0e3c8a182e1297b40cefdc0f272ecdff23ba53df --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/2825bc34-e49a-4342-8590-e899d26db9b6.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-sonnet-4-5-september-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4.5 (September 2025)", + "id": "anthropic/claude-sonnet-4-5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4.5 (September 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "75.6" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0709, + "details": { + "accuracy_raw": "7.09%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "12", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "75.6" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/951f55a7-6e78-4d43-a702-bbe327dc275f.json b/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/951f55a7-6e78-4d43-a702-bbe327dc275f.json new file mode 100644 index 0000000000000000000000000000000000000000..2daaef90657905ca24a81702418f3f5f64aed772 --- /dev/null +++ b/data/hal/hal-assistantbench/anthropic/claude-sonnet-4-5/951f55a7-6e78-4d43-a702-bbe327dc275f.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/claude-sonnet-4-5-high-september-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4.5 High (September 2025)", + "id": "anthropic/claude-sonnet-4-5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4.5 High (September 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "high", + "total_cost_usd": "99.23" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.118, + "details": { + "accuracy_raw": "11.80%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "9", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "99.23" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/deepseek/deepseek-r1/46a91d83-a641-465e-bc6b-4a007285aa19.json b/data/hal/hal-assistantbench/deepseek/deepseek-r1/46a91d83-a641-465e-bc6b-4a007285aa19.json new file mode 100644 index 0000000000000000000000000000000000000000..b55394ec729b98329c5416cdb17c2e5b47862595 --- /dev/null +++ b/data/hal/hal-assistantbench/deepseek/deepseek-r1/46a91d83-a641-465e-bc6b-4a007285aa19.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/deepseek-r1-january-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek R1 (January 2025)", + "id": "deepseek/deepseek-r1", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek R1 (January 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "5.3" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0, + "details": { + "accuracy_raw": "0.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "15", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "5.3" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/deepseek/deepseek-r1/c0182cb5-aaa4-4900-b32e-c5a8931d75de.json b/data/hal/hal-assistantbench/deepseek/deepseek-r1/c0182cb5-aaa4-4900-b32e-c5a8931d75de.json new file mode 100644 index 0000000000000000000000000000000000000000..ab6b7abfc90403900141b8d118f99bae578391f7 --- /dev/null +++ b/data/hal/hal-assistantbench/deepseek/deepseek-r1/c0182cb5-aaa4-4900-b32e-c5a8931d75de.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/deepseek-r1-may-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek R1 (May 2025)", + "id": "deepseek/deepseek-r1", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek R1 (May 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "18.18" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0875, + "details": { + "accuracy_raw": "8.75%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "10", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "18.18" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/deepseek/deepseek-v3/10ed5bf5-d4c1-445a-a169-feaafb9b6e0d.json b/data/hal/hal-assistantbench/deepseek/deepseek-v3/10ed5bf5-d4c1-445a-a169-feaafb9b6e0d.json new file mode 100644 index 0000000000000000000000000000000000000000..0521eb6ff0189fc2393f54c8cd8c2730db61b82d --- /dev/null +++ b/data/hal/hal-assistantbench/deepseek/deepseek-v3/10ed5bf5-d4c1-445a-a169-feaafb9b6e0d.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/deepseek-v3-march-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek V3 (March 2025)", + "id": "deepseek/deepseek-v3", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek V3 (March 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "12.66" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0203, + "details": { + "accuracy_raw": "2.03%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "14", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "12.66" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/google/gemini-2-0-flash/718e339e-3965-42df-9b62-8b8c3881abb9.json b/data/hal/hal-assistantbench/google/gemini-2-0-flash/718e339e-3965-42df-9b62-8b8c3881abb9.json new file mode 100644 index 0000000000000000000000000000000000000000..4a11841607368d65a73861b7d243e7cd436a3700 --- /dev/null +++ b/data/hal/hal-assistantbench/google/gemini-2-0-flash/718e339e-3965-42df-9b62-8b8c3881abb9.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/gemini-2-0-flash-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Gemini 2.0 Flash (February 2025)", + "id": "google/gemini-2.0-flash", + "developer": "google", + "additional_details": { + "hal_model_name": "Gemini 2.0 Flash (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "2.18" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0262, + "details": { + "accuracy_raw": "2.62%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "13", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "2.18" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/openai/gpt-4-1/64d0a36e-d5f0-4482-9b99-26593fe0c6fa.json b/data/hal/hal-assistantbench/openai/gpt-4-1/64d0a36e-d5f0-4482-9b99-26593fe0c6fa.json new file mode 100644 index 0000000000000000000000000000000000000000..5eafd78df500993844b3ce8b446a5058258d6364 --- /dev/null +++ b/data/hal/hal-assistantbench/openai/gpt-4-1/64d0a36e-d5f0-4482-9b99-26593fe0c6fa.json @@ -0,0 +1,87 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "total_cost_usd": "14.15" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1739, + "details": { + "accuracy_raw": "17.39%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "5", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "14.15" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/openai/gpt-5/4d06020c-b2e8-4d36-987e-03420fa51068.json b/data/hal/hal-assistantbench/openai/gpt-5/4d06020c-b2e8-4d36-987e-03420fa51068.json new file mode 100644 index 0000000000000000000000000000000000000000..40700c7f3c906c6df068fa7bd07e2ed0f04d700f --- /dev/null +++ b/data/hal/hal-assistantbench/openai/gpt-5/4d06020c-b2e8-4d36-987e-03420fa51068.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/gpt-5-medium-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-5 Medium (August 2025)", + "id": "openai/gpt-5", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-5 Medium (August 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "medium", + "total_cost_usd": "41.69" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3523, + "details": { + "accuracy_raw": "35.23%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "2", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "41.69" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/openai/o3/38f886cf-f18b-44bb-914b-de6db1f49ee2.json b/data/hal/hal-assistantbench/openai/o3/38f886cf-f18b-44bb-914b-de6db1f49ee2.json new file mode 100644 index 0000000000000000000000000000000000000000..2606aa815c7eed2fbd0b5dacfe8989cca65a2d72 --- /dev/null +++ b/data/hal/hal-assistantbench/openai/o3/38f886cf-f18b-44bb-914b-de6db1f49ee2.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "medium", + "total_cost_usd": "15.15" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3881, + "details": { + "accuracy_raw": "38.81%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "1", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "15.15" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/openai/o4-mini/4017f840-f14a-4be3-8645-8816696b3dab.json b/data/hal/hal-assistantbench/openai/o4-mini/4017f840-f14a-4be3-8645-8816696b3dab.json new file mode 100644 index 0000000000000000000000000000000000000000..cc932bd5e89fc97a20b07adf7077d83062147d40 --- /dev/null +++ b/data/hal/hal-assistantbench/openai/o4-mini/4017f840-f14a-4be3-8645-8816696b3dab.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "low", + "total_cost_usd": "9.22" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2805, + "details": { + "accuracy_raw": "28.05%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "3", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "9.22" + } + } + } + ] +} diff --git a/data/hal/hal-assistantbench/openai/o4-mini/f23ca46f-e99a-4314-853a-caf4590c49b9.json b/data/hal/hal-assistantbench/openai/o4-mini/f23ca46f-e99a-4314-853a-caf4590c49b9.json new file mode 100644 index 0000000000000000000000000000000000000000..e87a0c73e5ae913d9462f790433f616224653fe1 --- /dev/null +++ b/data/hal/hal-assistantbench/openai/o4-mini/f23ca46f-e99a-4314-853a-caf4590c49b9.json @@ -0,0 +1,88 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-assistantbench/browser-use/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — AssistantBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "assistantbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "AssistantBench", + "inference_effort": "high", + "total_cost_usd": "16.39" + } + }, + "evaluation_results": [ + { + "evaluation_name": "AssistantBench", + "source_data": { + "source_type": "url", + "dataset_name": "AssistantBench", + "url": [ + "https://assistantbench.github.io", + "https://hal.cs.princeton.edu/assistantbench" + ], + "additional_details": { + "tasks_evaluated": "33", + "full_benchmark_size": "214", + "note": "HAL evaluates on a 33-task subset; full AssistantBench has 214 tasks" + } + }, + "metric_config": { + "evaluation_description": "Accuracy on HAL's 33-task AssistantBench split (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2384, + "details": { + "accuracy_raw": "23.84%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "web_search", + "description": "Search the web for information" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "4", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "16.39" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/214ad68c-266c-4dca-bb6a-884fffff3983.json b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/214ad68c-266c-4dca-bb6a-884fffff3983.json new file mode 100644 index 0000000000000000000000000000000000000000..ff52e9bedca37f8b3c3a68b635d7112b955846ea --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/214ad68c-266c-4dca-bb6a-884fffff3983.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "total_cost_usd": "291.97" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2833, + "details": { + "accuracy_raw": "28.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "18", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "291.97" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/47f2d34e-450b-49ec-8d4a-432521d6fc56.json b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/47f2d34e-450b-49ec-8d4a-432521d6fc56.json new file mode 100644 index 0000000000000000000000000000000000000000..478870012c065e7eea355bc4c3938b980bdc3636 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/47f2d34e-450b-49ec-8d4a-432521d6fc56.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "1151.88" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3933, + "details": { + "accuracy_raw": "39.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "3", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "1151.88" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/7a8d5b3c-3798-49cc-a36e-77e464a0a112.json b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/7a8d5b3c-3798-49cc-a36e-77e464a0a112.json new file mode 100644 index 0000000000000000000000000000000000000000..8b974cbb902e8b59a28dfa7e8cc98f6f91abb9e7 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/7a8d5b3c-3798-49cc-a36e-77e464a0a112.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "926.48" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3833, + "details": { + "accuracy_raw": "38.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "6", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "926.48" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/c3a274a4-b441-459b-b4d2-707adcab1e5f.json b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/c3a274a4-b441-459b-b4d2-707adcab1e5f.json new file mode 100644 index 0000000000000000000000000000000000000000..4b594463b41e92deeeb02b0bd60df64f374fa853 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-3-7-sonnet-20250219/c3a274a4-b441-459b-b4d2-707adcab1e5f.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "367.51" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3033, + "details": { + "accuracy_raw": "30.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "15", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "367.51" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/474581cc-60a7-479a-9bdb-4680c880a07c.json b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/474581cc-60a7-479a-9bdb-4680c880a07c.json new file mode 100644 index 0000000000000000000000000000000000000000..c00dddfc00c3dde1d5eac4d40974b119166f2677 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/474581cc-60a7-479a-9bdb-4680c880a07c.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/claude-sonnet-4-high-may-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4 High (May 2025)", + "id": "anthropic/claude-sonnet-4", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4 High (May 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "326.41" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3667, + "details": { + "accuracy_raw": "36.67%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "8", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "326.41" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/5f3dea3e-715e-406d-b6f6-1991cb9f4aa5.json b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/5f3dea3e-715e-406d-b6f6-1991cb9f4aa5.json new file mode 100644 index 0000000000000000000000000000000000000000..f3c9dfbfe2cdf46a1462c53f6f71014f09348718 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/5f3dea3e-715e-406d-b6f6-1991cb9f4aa5.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/claude-sonnet-4-may-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4 (May 2025)", + "id": "anthropic/claude-sonnet-4", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4 (May 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "1577.26" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4, + "details": { + "accuracy_raw": "40.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "2", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "1577.26" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/8189a583-1733-4cf4-98a7-636303b72330.json b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/8189a583-1733-4cf4-98a7-636303b72330.json new file mode 100644 index 0000000000000000000000000000000000000000..d65df1f4a8a52087cdc0717c6656db3214063bc5 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/8189a583-1733-4cf4-98a7-636303b72330.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/claude-sonnet-4-high-may-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4 High (May 2025)", + "id": "anthropic/claude-sonnet-4", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4 High (May 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "1609.92" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3933, + "details": { + "accuracy_raw": "39.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "4", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "1609.92" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/e09256f4-f3f9-46de-9c82-fb30a9cb39c9.json b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/e09256f4-f3f9-46de-9c82-fb30a9cb39c9.json new file mode 100644 index 0000000000000000000000000000000000000000..b4ff5000d0d66aa779c9737afe14a8d0422abe56 --- /dev/null +++ b/data/hal/hal-online-mind2web/anthropic/claude-sonnet-4/e09256f4-f3f9-46de-9c82-fb30a9cb39c9.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/claude-sonnet-4-may-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4 (May 2025)", + "id": "anthropic/claude-sonnet-4", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4 (May 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "total_cost_usd": "246.18" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3667, + "details": { + "accuracy_raw": "36.67%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "7", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "246.18" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/deepseek/deepseek-r1/d01fb056-4520-4ed3-9b69-310bceb9f50a.json b/data/hal/hal-online-mind2web/deepseek/deepseek-r1/d01fb056-4520-4ed3-9b69-310bceb9f50a.json new file mode 100644 index 0000000000000000000000000000000000000000..c8153f1ea15679258e876ae7041171766487eca7 --- /dev/null +++ b/data/hal/hal-online-mind2web/deepseek/deepseek-r1/d01fb056-4520-4ed3-9b69-310bceb9f50a.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/deepseek-r1-january-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek R1 (January 2025)", + "id": "deepseek/deepseek-r1", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek R1 (January 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "280.93" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2533, + "details": { + "accuracy_raw": "25.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "20", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "280.93" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/deepseek/deepseek-v3/bfea7e2a-311a-4df4-9a0f-01357ac2c7de.json b/data/hal/hal-online-mind2web/deepseek/deepseek-v3/bfea7e2a-311a-4df4-9a0f-01357ac2c7de.json new file mode 100644 index 0000000000000000000000000000000000000000..4df3c0de1aed87a4ac6b8b05813252449c4e19ad --- /dev/null +++ b/data/hal/hal-online-mind2web/deepseek/deepseek-v3/bfea7e2a-311a-4df4-9a0f-01357ac2c7de.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/deepseek-v3-march-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek V3 (March 2025)", + "id": "deepseek/deepseek-v3", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek V3 (March 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "214.74" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3233, + "details": { + "accuracy_raw": "32.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "10", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "214.74" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/google/gemini-2-0-flash/73c40907-a14d-4553-952b-404454e545c7.json b/data/hal/hal-online-mind2web/google/gemini-2-0-flash/73c40907-a14d-4553-952b-404454e545c7.json new file mode 100644 index 0000000000000000000000000000000000000000..a58101556720bc03011b384b70bd105e1e47da38 --- /dev/null +++ b/data/hal/hal-online-mind2web/google/gemini-2-0-flash/73c40907-a14d-4553-952b-404454e545c7.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/gemini-2-0-flash-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Gemini 2.0 Flash (February 2025)", + "id": "google/gemini-2.0-flash", + "developer": "google", + "additional_details": { + "hal_model_name": "Gemini 2.0 Flash (February 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "8.83" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.29, + "details": { + "accuracy_raw": "29.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "16", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "8.83" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/google/gemini-2-0-flash/7c042687-67e6-44f5-9d07-a623a391310c.json b/data/hal/hal-online-mind2web/google/gemini-2-0-flash/7c042687-67e6-44f5-9d07-a623a391310c.json new file mode 100644 index 0000000000000000000000000000000000000000..2ccdeff4b7a9bfbb1b733cbcd0710269c533e7db --- /dev/null +++ b/data/hal/hal-online-mind2web/google/gemini-2-0-flash/7c042687-67e6-44f5-9d07-a623a391310c.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/gemini-2-0-flash-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Gemini 2.0 Flash (February 2025)", + "id": "google/gemini-2.0-flash", + "developer": "google", + "additional_details": { + "hal_model_name": "Gemini 2.0 Flash (February 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "total_cost_usd": "5.03" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2667, + "details": { + "accuracy_raw": "26.67%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "19", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "5.03" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/gpt-4-1/47244e96-bbe1-4528-b8d7-08db6950580c.json b/data/hal/hal-online-mind2web/openai/gpt-4-1/47244e96-bbe1-4528-b8d7-08db6950580c.json new file mode 100644 index 0000000000000000000000000000000000000000..1ecc40e6296d62e52cae97a4bbb1fe35bacdfc1e --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/gpt-4-1/47244e96-bbe1-4528-b8d7-08db6950580c.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "total_cost_usd": "271.24" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3033, + "details": { + "accuracy_raw": "30.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "14", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "271.24" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/gpt-4-1/fa50eb2b-59eb-44c5-b956-5173236056aa.json b/data/hal/hal-online-mind2web/openai/gpt-4-1/fa50eb2b-59eb-44c5-b956-5173236056aa.json new file mode 100644 index 0000000000000000000000000000000000000000..6ca4c4c048eba23559350bc1b6faf15f14264d46 --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/gpt-4-1/fa50eb2b-59eb-44c5-b956-5173236056aa.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "total_cost_usd": "236.62" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3633, + "details": { + "accuracy_raw": "36.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "9", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "236.62" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/gpt-5/6b2a0c38-77e3-42e7-87cd-775a59166de9.json b/data/hal/hal-online-mind2web/openai/gpt-5/6b2a0c38-77e3-42e7-87cd-775a59166de9.json new file mode 100644 index 0000000000000000000000000000000000000000..39bdb36e0c63053b27141ea7beb57d888b6453a0 --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/gpt-5/6b2a0c38-77e3-42e7-87cd-775a59166de9.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/gpt-5-medium-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-5 Medium (August 2025)", + "id": "openai/gpt-5", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-5 Medium (August 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "medium", + "total_cost_usd": "171.07" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4233, + "details": { + "accuracy_raw": "42.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "1", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "171.07" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/gpt-5/eef13529-2edf-482d-9d65-fbcaa781f2b7.json b/data/hal/hal-online-mind2web/openai/gpt-5/eef13529-2edf-482d-9d65-fbcaa781f2b7.json new file mode 100644 index 0000000000000000000000000000000000000000..db3f88d71a6b5cdaceb9686fbcb9c288ed48f6dc --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/gpt-5/eef13529-2edf-482d-9d65-fbcaa781f2b7.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/gpt-5-medium-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-5 Medium (August 2025)", + "id": "openai/gpt-5", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-5 Medium (August 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "medium", + "total_cost_usd": "736.31" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.32, + "details": { + "accuracy_raw": "32.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "12", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "736.31" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o3/3e9c1c04-629d-4348-b529-819444161313.json b/data/hal/hal-online-mind2web/openai/o3/3e9c1c04-629d-4348-b529-819444161313.json new file mode 100644 index 0000000000000000000000000000000000000000..619d197945d9446fa8912afbaac8bb5d5ed5b99d --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o3/3e9c1c04-629d-4348-b529-819444161313.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "medium", + "total_cost_usd": "258.74" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.39, + "details": { + "accuracy_raw": "39.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "5", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "258.74" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o3/bf1bf80f-8375-49a9-ac90-4e67b1c4ea51.json b/data/hal/hal-online-mind2web/openai/o3/bf1bf80f-8375-49a9-ac90-4e67b1c4ea51.json new file mode 100644 index 0000000000000000000000000000000000000000..63c20a35cad4e58f965868ca4a4bfa91304e7695 --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o3/bf1bf80f-8375-49a9-ac90-4e67b1c4ea51.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "medium", + "total_cost_usd": "371.59" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.29, + "details": { + "accuracy_raw": "29.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "17", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "371.59" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o4-mini/408e6d84-9c17-42f6-b5c0-3c9fa9c7385f.json b/data/hal/hal-online-mind2web/openai/o4-mini/408e6d84-9c17-42f6-b5c0-3c9fa9c7385f.json new file mode 100644 index 0000000000000000000000000000000000000000..2e3ec618d48f4ddb799a0f00cbaac8b81302568b --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o4-mini/408e6d84-9c17-42f6-b5c0-3c9fa9c7385f.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "low", + "total_cost_usd": "162.36" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3167, + "details": { + "accuracy_raw": "31.67%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "13", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "162.36" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o4-mini/842f7023-d7d9-44b1-bcce-df01b8d501d7.json b/data/hal/hal-online-mind2web/openai/o4-mini/842f7023-d7d9-44b1-bcce-df01b8d501d7.json new file mode 100644 index 0000000000000000000000000000000000000000..b5d453fde205943b9c37b088289d198aac8300f4 --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o4-mini/842f7023-d7d9-44b1-bcce-df01b8d501d7.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "297.93" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2, + "details": { + "accuracy_raw": "20.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "21", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "297.93" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o4-mini/aa8e3048-40c2-45af-821d-b61dca8ec214.json b/data/hal/hal-online-mind2web/openai/o4-mini/aa8e3048-40c2-45af-821d-b61dca8ec214.json new file mode 100644 index 0000000000000000000000000000000000000000..a9eefe7c7b6c185739ff799ae6d8942f7640b21d --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o4-mini/aa8e3048-40c2-45af-821d-b61dca8ec214.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/browser-use/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "Browser-Use", + "benchmark": "Online Mind2Web", + "inference_effort": "low", + "total_cost_usd": "201.44" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1833, + "details": { + "accuracy_raw": "18.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "Browser-Use", + "hal_rank": "22", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "201.44" + } + } + } + ] +} diff --git a/data/hal/hal-online-mind2web/openai/o4-mini/d79e8f9c-c40c-453f-9e1f-1f5ad2de07a8.json b/data/hal/hal-online-mind2web/openai/o4-mini/d79e8f9c-c40c-453f-9e1f-1f5ad2de07a8.json new file mode 100644 index 0000000000000000000000000000000000000000..39f58926dcbf0d0ca7cf586d4f0d0084e0fa4784 --- /dev/null +++ b/data/hal/hal-online-mind2web/openai/o4-mini/d79e8f9c-c40c-453f-9e1f-1f5ad2de07a8.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-online-mind2web/seeact/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — Online Mind2Web", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Web Assistance", + "benchmark_slug": "online_mind2web" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "SeeAct", + "benchmark": "Online Mind2Web", + "inference_effort": "high", + "total_cost_usd": "228.98" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Online Mind2Web", + "source_data": { + "source_type": "url", + "dataset_name": "Online Mind2Web", + "url": [ + "https://osu-nlp-group.github.io/Mind2Web/", + "https://hal.cs.princeton.edu/online_mind2web" + ] + }, + "metric_config": { + "evaluation_description": "Task success rate on Online Mind2Web (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.32, + "details": { + "accuracy_raw": "32.00%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "browser", + "description": "Navigate and interact with live web pages" + }, + { + "name": "click", + "description": "Click on web page elements" + }, + { + "name": "type", + "description": "Type text into web page inputs" + }, + { + "name": "scroll", + "description": "Scroll web pages" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SeeAct", + "hal_rank": "11", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "228.98" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/0e8dde2b-97f9-4e09-ba13-4c30d4aff06c.json b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/0e8dde2b-97f9-4e09-ba13-4c30d4aff06c.json new file mode 100644 index 0000000000000000000000000000000000000000..70c38bc24bf8fbf39ad65e0a302545d0bcb7a804 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/0e8dde2b-97f9-4e09-ba13-4c30d4aff06c.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "48.28" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1765, + "details": { + "accuracy_raw": "17.65%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "17", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "48.28" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/15d5ca6c-85c8-4835-8b8b-61fc3b046c98.json b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/15d5ca6c-85c8-4835-8b8b-61fc3b046c98.json new file mode 100644 index 0000000000000000000000000000000000000000..92e47391534267528c5269927d0a4604f975bbb8 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/15d5ca6c-85c8-4835-8b8b-61fc3b046c98.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "7.12" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2255, + "details": { + "accuracy_raw": "22.55%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "13", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "7.12" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/18b685af-979f-42da-bd56-eb1013ee2117.json b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/18b685af-979f-42da-bd56-eb1013ee2117.json new file mode 100644 index 0000000000000000000000000000000000000000..b074f7742c4c072eba6b2e1a235a9b410a9c9122 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/18b685af-979f-42da-bd56-eb1013ee2117.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "11.74" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3039, + "details": { + "accuracy_raw": "30.39%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "3", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "11.74" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/480f7800-8c7b-4230-a0c1-8e81fb3d61b9.json b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/480f7800-8c7b-4230-a0c1-8e81fb3d61b9.json new file mode 100644 index 0000000000000000000000000000000000000000..6b41495c5ce072be1f3b0d782d883d312c02ce31 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-3-7-sonnet-20250219/480f7800-8c7b-4230-a0c1-8e81fb3d61b9.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "41.22" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1078, + "details": { + "accuracy_raw": "10.78%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "20", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "41.22" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/38acb0a5-1384-4f3c-8fd4-61ee2a5a9013.json b/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/38acb0a5-1384-4f3c-8fd4-61ee2a5a9013.json new file mode 100644 index 0000000000000000000000000000000000000000..2939c4d33dcb6779179c057ea3900f6f07490ae5 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/38acb0a5-1384-4f3c-8fd4-61ee2a5a9013.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-haiku-4-5-high-october-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Haiku 4.5 High (October 2025)", + "id": "anthropic/claude-haiku-4.5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Haiku 4.5 High (October 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "3.41" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2353, + "details": { + "accuracy_raw": "23.53%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "11", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "3.41" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/d3d9eafb-b8e5-4168-82b0-de266ed01c3a.json b/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/d3d9eafb-b8e5-4168-82b0-de266ed01c3a.json new file mode 100644 index 0000000000000000000000000000000000000000..2a11a44b6c63e621ec8218a4c20747e16b7e98de --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-haiku-4-5/d3d9eafb-b8e5-4168-82b0-de266ed01c3a.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-haiku-4-5-october-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Haiku 4.5 (October 2025)", + "id": "anthropic/claude-haiku-4.5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Haiku 4.5 (October 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "2.66" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1863, + "details": { + "accuracy_raw": "18.63%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "16", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "2.66" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/3a2abea6-5dd5-4dd8-937e-763792ca811f.json b/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/3a2abea6-5dd5-4dd8-937e-763792ca811f.json new file mode 100644 index 0000000000000000000000000000000000000000..61fd5cc06158cbc2b7910392b84ece6919c00fdc --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/3a2abea6-5dd5-4dd8-937e-763792ca811f.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-opus-4-1-high-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 High (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 High (August 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "33.75" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2647, + "details": { + "accuracy_raw": "26.47%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "9", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "33.75" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/904a889d-bbaa-49cd-b77d-96265cfafd55.json b/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/904a889d-bbaa-49cd-b77d-96265cfafd55.json new file mode 100644 index 0000000000000000000000000000000000000000..f474955a717f4f22989b8513b9ec5677a9958f68 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-opus-4-1/904a889d-bbaa-49cd-b77d-96265cfafd55.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-opus-4-1-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 (August 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "33.37" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2745, + "details": { + "accuracy_raw": "27.45%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "8", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "33.37" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/811d5fb5-4ce3-483a-a25d-4f0854746e01.json b/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/811d5fb5-4ce3-483a-a25d-4f0854746e01.json new file mode 100644 index 0000000000000000000000000000000000000000..ebd1db6fa79a0d3c95010bd6eff1a789106c42ca --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/811d5fb5-4ce3-483a-a25d-4f0854746e01.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-sonnet-4-5-high-september-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4.5 High (September 2025)", + "id": "anthropic/claude-sonnet-4-5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4.5 High (September 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "7.47" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3039, + "details": { + "accuracy_raw": "30.39%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "2", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "7.47" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/ab869ad5-b266-4209-b5bc-4a7b881042bc.json b/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/ab869ad5-b266-4209-b5bc-4a7b881042bc.json new file mode 100644 index 0000000000000000000000000000000000000000..a9d2dc69ed5183a095e275230edf4c08bf6ac1a9 --- /dev/null +++ b/data/hal/hal-scienceagentbench/anthropic/claude-sonnet-4-5/ab869ad5-b266-4209-b5bc-4a7b881042bc.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/claude-sonnet-4-5-september-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Sonnet 4.5 (September 2025)", + "id": "anthropic/claude-sonnet-4-5", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Sonnet 4.5 (September 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "7.39" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2941, + "details": { + "accuracy_raw": "29.41%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "5", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "7.39" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/deepseek/deepseek-r1/62bfa937-e1bd-478c-81ce-c0a279db8d19.json b/data/hal/hal-scienceagentbench/deepseek/deepseek-r1/62bfa937-e1bd-478c-81ce-c0a279db8d19.json new file mode 100644 index 0000000000000000000000000000000000000000..a9c748874b250cedc5f2cee6f81a551609204fef --- /dev/null +++ b/data/hal/hal-scienceagentbench/deepseek/deepseek-r1/62bfa937-e1bd-478c-81ce-c0a279db8d19.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/deepseek-r1-january-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek R1 (January 2025)", + "id": "deepseek/deepseek-r1", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek R1 (January 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "18.24" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2353, + "details": { + "accuracy_raw": "23.53%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "12", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "18.24" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/86fd6af6-2995-4648-9015-421341e5f0f9.json b/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/86fd6af6-2995-4648-9015-421341e5f0f9.json new file mode 100644 index 0000000000000000000000000000000000000000..2f2aacd9e510d23ddeaf3bd59e5b75db38c6ba94 --- /dev/null +++ b/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/86fd6af6-2995-4648-9015-421341e5f0f9.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/deepseek-v3-march-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek V3 (March 2025)", + "id": "deepseek/deepseek-v3", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek V3 (March 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "2.09" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1569, + "details": { + "accuracy_raw": "15.69%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "18", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "2.09" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/f276f8b4-3daf-485d-ac28-3955b0f3df51.json b/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/f276f8b4-3daf-485d-ac28-3955b0f3df51.json new file mode 100644 index 0000000000000000000000000000000000000000..0790d30ebbd2482dbf1a87ed9cbfec66c19a7a94 --- /dev/null +++ b/data/hal/hal-scienceagentbench/deepseek/deepseek-v3/f276f8b4-3daf-485d-ac28-3955b0f3df51.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/deepseek-v3-march-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek V3 (March 2025)", + "id": "deepseek/deepseek-v3", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek V3 (March 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "55.73" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0098, + "details": { + "accuracy_raw": "0.98%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "23", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "55.73" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/google/gemini-2-0-flash/7f5e6bc1-2397-433a-bbf2-72b5c72ce3a1.json b/data/hal/hal-scienceagentbench/google/gemini-2-0-flash/7f5e6bc1-2397-433a-bbf2-72b5c72ce3a1.json new file mode 100644 index 0000000000000000000000000000000000000000..52c38fd0d587576104e24eee1b1dd8365b6de621 --- /dev/null +++ b/data/hal/hal-scienceagentbench/google/gemini-2-0-flash/7f5e6bc1-2397-433a-bbf2-72b5c72ce3a1.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/gemini-2-0-flash-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Gemini 2.0 Flash (February 2025)", + "id": "google/gemini-2.0-flash", + "developer": "google", + "additional_details": { + "hal_model_name": "Gemini 2.0 Flash (February 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "0.19" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1275, + "details": { + "accuracy_raw": "12.75%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "19", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "0.19" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/gpt-4-1/6e2565f4-d139-41cd-ba01-eb1ff3e7eb4f.json b/data/hal/hal-scienceagentbench/openai/gpt-4-1/6e2565f4-d139-41cd-ba01-eb1ff3e7eb4f.json new file mode 100644 index 0000000000000000000000000000000000000000..18adaf9f1414ae9b0d492553aad5f630b9a48f4c --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/gpt-4-1/6e2565f4-d139-41cd-ba01-eb1ff3e7eb4f.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "68.95" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.0686, + "details": { + "accuracy_raw": "6.86%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "22", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "68.95" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/gpt-4-1/87fb8284-ae7e-4c29-a89a-572029ea14f0.json b/data/hal/hal-scienceagentbench/openai/gpt-4-1/87fb8284-ae7e-4c29-a89a-572029ea14f0.json new file mode 100644 index 0000000000000000000000000000000000000000..8701454a04b564fab720cfe69338a7309a9a97af --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/gpt-4-1/87fb8284-ae7e-4c29-a89a-572029ea14f0.json @@ -0,0 +1,90 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "total_cost_usd": "7.42" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2451, + "details": { + "accuracy_raw": "24.51%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "10", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "7.42" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/gpt-5/4e5ba185-0826-44d2-91eb-9cd8b2661acf.json b/data/hal/hal-scienceagentbench/openai/gpt-5/4e5ba185-0826-44d2-91eb-9cd8b2661acf.json new file mode 100644 index 0000000000000000000000000000000000000000..76c61027a0079851392668a9114737e647dc0e8e --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/gpt-5/4e5ba185-0826-44d2-91eb-9cd8b2661acf.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/gpt-5-medium-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-5 Medium (August 2025)", + "id": "openai/gpt-5", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-5 Medium (August 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "medium", + "total_cost_usd": "18.26" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3039, + "details": { + "accuracy_raw": "30.39%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "4", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "18.26" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o3/75395466-2c7f-4026-bbbc-51217fd08cb8.json b/data/hal/hal-scienceagentbench/openai/o3/75395466-2c7f-4026-bbbc-51217fd08cb8.json new file mode 100644 index 0000000000000000000000000000000000000000..a3c5f15e4531f39b6ca8c8dd889dc5d57901608d --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o3/75395466-2c7f-4026-bbbc-51217fd08cb8.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "inference_effort": "medium", + "total_cost_usd": "31.08" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.098, + "details": { + "accuracy_raw": "9.80%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "21", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "31.08" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o3/c0cdc368-4f88-40c2-bb59-d0fcb54f3edd.json b/data/hal/hal-scienceagentbench/openai/o3/c0cdc368-4f88-40c2-bb59-d0fcb54f3edd.json new file mode 100644 index 0000000000000000000000000000000000000000..6bfd92c00ee1a423bc2d9f50b795ebdb221bd2eb --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o3/c0cdc368-4f88-40c2-bb59-d0fcb54f3edd.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "medium", + "total_cost_usd": "11.69" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3333, + "details": { + "accuracy_raw": "33.33%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "1", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "11.69" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o4-mini/6d07d448-f549-4965-8ff0-b46230050bb2.json b/data/hal/hal-scienceagentbench/openai/o4-mini/6d07d448-f549-4965-8ff0-b46230050bb2.json new file mode 100644 index 0000000000000000000000000000000000000000..30e8c23f33a4b77b469ff3625d9859f1b061d8d6 --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o4-mini/6d07d448-f549-4965-8ff0-b46230050bb2.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "low", + "total_cost_usd": "3.95" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2745, + "details": { + "accuracy_raw": "27.45%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "6", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "3.95" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o4-mini/a15188b4-147a-4400-b499-824e1aa02ae6.json b/data/hal/hal-scienceagentbench/openai/o4-mini/a15188b4-147a-4400-b499-824e1aa02ae6.json new file mode 100644 index 0000000000000000000000000000000000000000..685307f9284b91898ef0a063fd5882a2fe8ddc43 --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o4-mini/a15188b4-147a-4400-b499-824e1aa02ae6.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "76.3" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2157, + "details": { + "accuracy_raw": "21.57%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "14", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "76.3" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o4-mini/c6b557af-15b9-4296-a706-add460a30148.json b/data/hal/hal-scienceagentbench/openai/o4-mini/c6b557af-15b9-4296-a706-add460a30148.json new file mode 100644 index 0000000000000000000000000000000000000000..46f1b10e78bcf28bb7818a2a84e5877881cb40a3 --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o4-mini/c6b557af-15b9-4296-a706-add460a30148.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/hal-generalist-agent/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "ScienceAgentBench", + "inference_effort": "low", + "total_cost_usd": "77.32" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1961, + "details": { + "accuracy_raw": "19.61%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "15", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "77.32" + } + } + } + ] +} diff --git a/data/hal/hal-scienceagentbench/openai/o4-mini/ccbbe473-86d0-4631-a221-feea3382480e.json b/data/hal/hal-scienceagentbench/openai/o4-mini/ccbbe473-86d0-4631-a221-feea3382480e.json new file mode 100644 index 0000000000000000000000000000000000000000..ad9e92108be3b253fa6c468b8eff4c9c1f84066d --- /dev/null +++ b/data/hal/hal-scienceagentbench/openai/o4-mini/ccbbe473-86d0-4631-a221-feea3382480e.json @@ -0,0 +1,91 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-scienceagentbench/sab-self-debug/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — ScienceAgentBench", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Scientific Programming", + "benchmark_slug": "scienceagentbench" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "SAB Self-Debug", + "benchmark": "ScienceAgentBench", + "inference_effort": "high", + "total_cost_usd": "11.18" + } + }, + "evaluation_results": [ + { + "evaluation_name": "ScienceAgentBench", + "source_data": { + "source_type": "url", + "dataset_name": "ScienceAgentBench", + "url": [ + "https://osu-nlp-group.github.io/ScienceAgentBench/", + "https://hal.cs.princeton.edu/scienceagentbench" + ] + }, + "metric_config": { + "evaluation_description": "Success rate on ScienceAgentBench tasks (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2745, + "details": { + "accuracy_raw": "27.45%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "python", + "description": "Execute Python code for data analysis" + }, + { + "name": "bash", + "description": "Execute shell commands" + }, + { + "name": "read_file", + "description": "Read datasets and files" + }, + { + "name": "write_file", + "description": "Write output files and results" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "SAB Self-Debug", + "hal_rank": "7", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "11.18" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/0c8f5d08-37d3-4629-aa55-20b4cc550fe5.json b/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/0c8f5d08-37d3-4629-aa55-20b4cc550fe5.json new file mode 100644 index 0000000000000000000000000000000000000000..322c5016996f29472c4e8fcca21b52a1d824d894 --- /dev/null +++ b/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/0c8f5d08-37d3-4629-aa55-20b4cc550fe5.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/claude-3-7-sonnet-high-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet High (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet High (February 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "high", + "total_cost_usd": "56.43" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2671, + "details": { + "accuracy_raw": "26.71%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "12", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "56.43" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/8364beb4-96cb-4503-8c95-adc68a8eb018.json b/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/8364beb4-96cb-4503-8c95-adc68a8eb018.json new file mode 100644 index 0000000000000000000000000000000000000000..d38f0627ebc597e39226929f2a28f36cbd4d2527 --- /dev/null +++ b/data/hal/hal-usaco/anthropic/claude-3-7-sonnet-20250219/8364beb4-96cb-4503-8c95-adc68a8eb018.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/claude-3-7-sonnet-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude-3.7 Sonnet (February 2025)", + "id": "anthropic/claude-3-7-sonnet-20250219", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude-3.7 Sonnet (February 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "38.7" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2932, + "details": { + "accuracy_raw": "29.32%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "10", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "38.7" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/anthropic/claude-opus-4-1/304327af-8650-4de4-9cf6-63864ef34e75.json b/data/hal/hal-usaco/anthropic/claude-opus-4-1/304327af-8650-4de4-9cf6-63864ef34e75.json new file mode 100644 index 0000000000000000000000000000000000000000..0065c88cc54f7fb5ef86ad9a60ad6a0a91f330b3 --- /dev/null +++ b/data/hal/hal-usaco/anthropic/claude-opus-4-1/304327af-8650-4de4-9cf6-63864ef34e75.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/claude-opus-4-1-high-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 High (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 High (August 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "high", + "total_cost_usd": "267.72" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5147, + "details": { + "accuracy_raw": "51.47%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "4", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "267.72" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/anthropic/claude-opus-4-1/3dc74364-1852-4d27-ba31-c7ab806e8b8e.json b/data/hal/hal-usaco/anthropic/claude-opus-4-1/3dc74364-1852-4d27-ba31-c7ab806e8b8e.json new file mode 100644 index 0000000000000000000000000000000000000000..a09a992d7b2a3aa9ab0a83d101ba9aa905ecb80c --- /dev/null +++ b/data/hal/hal-usaco/anthropic/claude-opus-4-1/3dc74364-1852-4d27-ba31-c7ab806e8b8e.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/claude-opus-4-1-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Claude Opus 4.1 (August 2025)", + "id": "anthropic/claude-opus-4-1", + "developer": "anthropic", + "additional_details": { + "hal_model_name": "Claude Opus 4.1 (August 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "276.19" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4821, + "details": { + "accuracy_raw": "48.21%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "5", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "276.19" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/deepseek/deepseek-r1/a337b879-aef9-490b-91c8-285cbfdf2e54.json b/data/hal/hal-usaco/deepseek/deepseek-r1/a337b879-aef9-490b-91c8-285cbfdf2e54.json new file mode 100644 index 0000000000000000000000000000000000000000..654cf3c087c5f19e1b3c51dff80909e53e5b34ab --- /dev/null +++ b/data/hal/hal-usaco/deepseek/deepseek-r1/a337b879-aef9-490b-91c8-285cbfdf2e54.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/deepseek-r1-january-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek R1 (January 2025)", + "id": "deepseek/deepseek-r1", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek R1 (January 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "80.04" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3811, + "details": { + "accuracy_raw": "38.11%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "9", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "80.04" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/deepseek/deepseek-v3/053d8d0e-4365-4726-aef7-487c72ae2a84.json b/data/hal/hal-usaco/deepseek/deepseek-v3/053d8d0e-4365-4726-aef7-487c72ae2a84.json new file mode 100644 index 0000000000000000000000000000000000000000..70504c473874bd409ab23411339462afc31d5b39 --- /dev/null +++ b/data/hal/hal-usaco/deepseek/deepseek-v3/053d8d0e-4365-4726-aef7-487c72ae2a84.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/deepseek-v3-march-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "DeepSeek V3 (March 2025)", + "id": "deepseek/deepseek-v3", + "developer": "deepseek", + "additional_details": { + "hal_model_name": "DeepSeek V3 (March 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "12.08" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3909, + "details": { + "accuracy_raw": "39.09%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "8", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "12.08" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/google/gemini-2-0-flash/7f144a3a-0b3d-449d-8d86-420577c9c7ec.json b/data/hal/hal-usaco/google/gemini-2-0-flash/7f144a3a-0b3d-449d-8d86-420577c9c7ec.json new file mode 100644 index 0000000000000000000000000000000000000000..5222b7308ace9c611499faa3d3eb9f551ab2efe9 --- /dev/null +++ b/data/hal/hal-usaco/google/gemini-2-0-flash/7f144a3a-0b3d-449d-8d86-420577c9c7ec.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/gemini-2-0-flash-february-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "Gemini 2.0 Flash (February 2025)", + "id": "google/gemini-2.0-flash", + "developer": "google", + "additional_details": { + "hal_model_name": "Gemini 2.0 Flash (February 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "1.46" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2704, + "details": { + "accuracy_raw": "27.04%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "11", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "1.46" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/gpt-4-1/52cfbd7a-e431-4259-ad35-75ef19b01345.json b/data/hal/hal-usaco/openai/gpt-4-1/52cfbd7a-e431-4259-ad35-75ef19b01345.json new file mode 100644 index 0000000000000000000000000000000000000000..945bfb9c29bee3c0117e7412e70ccd7e75e2a3d7 --- /dev/null +++ b/data/hal/hal-usaco/openai/gpt-4-1/52cfbd7a-e431-4259-ad35-75ef19b01345.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/hal-generalist-agent/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "HAL Generalist Agent", + "benchmark": "USACO", + "total_cost_usd": "197.33" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.2541, + "details": { + "accuracy_raw": "25.41%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "HAL Generalist Agent", + "hal_rank": "13", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "197.33" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/gpt-4-1/f8a90b9e-42f3-4029-aceb-1287cbbfe1ba.json b/data/hal/hal-usaco/openai/gpt-4-1/f8a90b9e-42f3-4029-aceb-1287cbbfe1ba.json new file mode 100644 index 0000000000000000000000000000000000000000..34489abb6f5ea2aab46c244eeff7e0669006db8d --- /dev/null +++ b/data/hal/hal-usaco/openai/gpt-4-1/f8a90b9e-42f3-4029-aceb-1287cbbfe1ba.json @@ -0,0 +1,82 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/gpt-4-1-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-4.1 (April 2025)", + "id": "openai/gpt-4.1", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-4.1 (April 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "total_cost_usd": "28.1" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4495, + "details": { + "accuracy_raw": "44.95%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "7", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "28.1" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/gpt-5/fc81e5bc-bb1b-487b-b485-872481dbb1b4.json b/data/hal/hal-usaco/openai/gpt-5/fc81e5bc-bb1b-487b-b485-872481dbb1b4.json new file mode 100644 index 0000000000000000000000000000000000000000..d3c22a2a4979269ce96ded76bc4f48f09a070b14 --- /dev/null +++ b/data/hal/hal-usaco/openai/gpt-5/fc81e5bc-bb1b-487b-b485-872481dbb1b4.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/gpt-5-medium-august-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "GPT-5 Medium (August 2025)", + "id": "openai/gpt-5", + "developer": "openai", + "additional_details": { + "hal_model_name": "GPT-5 Medium (August 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "medium", + "total_cost_usd": "116.63" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6906, + "details": { + "accuracy_raw": "69.06%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "1", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "116.63" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/o3/036f31c0-5f4d-4509-a918-18d3ae25059a.json b/data/hal/hal-usaco/openai/o3/036f31c0-5f4d-4509-a918-18d3ae25059a.json new file mode 100644 index 0000000000000000000000000000000000000000..8dfb8a0996f3d71d77ff3695932243fa9b8a5ab5 --- /dev/null +++ b/data/hal/hal-usaco/openai/o3/036f31c0-5f4d-4509-a918-18d3ae25059a.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/o3-medium-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o3 Medium (April 2025)", + "id": "openai/o3", + "developer": "openai", + "additional_details": { + "hal_model_name": "o3 Medium (April 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "medium", + "total_cost_usd": "57.3" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4625, + "details": { + "accuracy_raw": "46.25%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "6", + "runs": "1", + "verified": "True", + "is_pareto": "False", + "total_cost_usd": "57.3" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/o4-mini/206d190c-8a0f-450a-bd70-6ab1a7d84d8a.json b/data/hal/hal-usaco/openai/o4-mini/206d190c-8a0f-450a-bd70-6ab1a7d84d8a.json new file mode 100644 index 0000000000000000000000000000000000000000..c562b9bfce8e362564a53a2fa1e7b8acdeaa9543 --- /dev/null +++ b/data/hal/hal-usaco/openai/o4-mini/206d190c-8a0f-450a-bd70-6ab1a7d84d8a.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/o4-mini-high-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini High (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini High (April 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "high", + "total_cost_usd": "77.28" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6482, + "details": { + "accuracy_raw": "64.82%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "2", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "77.28" + } + } + } + ] +} diff --git a/data/hal/hal-usaco/openai/o4-mini/700e3e02-bffa-49ed-bcbb-4e700b541ba8.json b/data/hal/hal-usaco/openai/o4-mini/700e3e02-bffa-49ed-bcbb-4e700b541ba8.json new file mode 100644 index 0000000000000000000000000000000000000000..c0ce0b793c1c911be3a7b8762e6470d2a5ac706a --- /dev/null +++ b/data/hal/hal-usaco/openai/o4-mini/700e3e02-bffa-49ed-bcbb-4e700b541ba8.json @@ -0,0 +1,83 @@ +{ + "schema_version": "0.2.2", + "evaluation_id": "hal-usaco/usaco-episodic-semantic/o4-mini-low-april-2025/1782803417.6616569", + "retrieved_timestamp": "1782803417.6616569", + "source_metadata": { + "source_name": "HAL Leaderboard — USACO", + "source_type": "documentation", + "source_organization_name": "Princeton SAgE Team", + "source_organization_url": "https://hal.cs.princeton.edu", + "evaluator_relationship": "third_party", + "additional_details": { + "paper": "https://arxiv.org/pdf/2510.11977", + "benchmark_category": "Programming", + "benchmark_slug": "usaco" + } + }, + "eval_library": { + "name": "HAL", + "version": "unknown" + }, + "model_info": { + "name": "o4-mini Low (April 2025)", + "id": "openai/o4-mini", + "developer": "openai", + "additional_details": { + "hal_model_name": "o4-mini Low (April 2025)", + "agent_scaffold": "USACO Episodic + Semantic", + "benchmark": "USACO", + "inference_effort": "low", + "total_cost_usd": "24.6" + } + }, + "evaluation_results": [ + { + "evaluation_name": "USACO", + "source_data": { + "source_type": "url", + "dataset_name": "USACO", + "url": [ + "https://usaco.guide", + "https://hal.cs.princeton.edu/usaco" + ] + }, + "metric_config": { + "evaluation_description": "Fraction of USACO problems solved (0.0–1.0)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5309, + "details": { + "accuracy_raw": "53.09%" + } + }, + "generation_config": { + "generation_args": { + "agentic_eval_config": { + "available_tools": [ + { + "name": "bash", + "description": "Execute shell commands and compile/run code" + }, + { + "name": "python", + "description": "Execute Python code" + } + ] + } + }, + "additional_details": { + "agent_scaffold": "USACO Episodic + Semantic", + "hal_rank": "3", + "runs": "1", + "verified": "True", + "is_pareto": "True", + "total_cost_usd": "24.6" + } + } + } + ] +} diff --git a/data/reward-bench/Ahjeong/MMPO_Gemma_7b/c395b556-978d-45f3-826e-b35807820958.json b/data/reward-bench/Ahjeong/MMPO_Gemma_7b/c395b556-978d-45f3-826e-b35807820958.json new file mode 100644 index 0000000000000000000000000000000000000000..d1b39e00685679c3de7c103ba58c1cf483375548 --- /dev/null +++ b/data/reward-bench/Ahjeong/MMPO_Gemma_7b/c395b556-978d-45f3-826e-b35807820958.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Ahjeong_MMPO_Gemma_7b/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Ahjeong/MMPO_Gemma_7b", + "id": "Ahjeong/MMPO_Gemma_7b", + "developer": "Ahjeong", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7587 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9693 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.614 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7135 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7756 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6831 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Ahjeong/MMPO_Gemma_7b_gamma1.1_epoch3/9ff0ecb8-ca76-4160-a807-015311b682fc.json b/data/reward-bench/Ahjeong/MMPO_Gemma_7b_gamma1.1_epoch3/9ff0ecb8-ca76-4160-a807-015311b682fc.json new file mode 100644 index 0000000000000000000000000000000000000000..7bb6e55182c12124fb1678b5ed9fd7c291a0d5e4 --- /dev/null +++ b/data/reward-bench/Ahjeong/MMPO_Gemma_7b_gamma1.1_epoch3/9ff0ecb8-ca76-4160-a807-015311b682fc.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Ahjeong_MMPO_Gemma_7b_gamma1.1_epoch3/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Ahjeong/MMPO_Gemma_7b_gamma1.1_epoch3", + "id": "Ahjeong/MMPO_Gemma_7b_gamma1.1_epoch3", + "developer": "Ahjeong", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7652 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9721 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6338 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7635 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7284 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6913 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/0ceffd81-8822-45ad-b04c-6d229dc395ed.json b/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/0ceffd81-8822-45ad-b04c-6d229dc395ed.json new file mode 100644 index 0000000000000000000000000000000000000000..b2c425eaf0151cc54362956522b5e3780ff51f87 --- /dev/null +++ b/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/0ceffd81-8822-45ad-b04c-6d229dc395ed.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/CIR-AMS_BTRM_Qwen2_7b_0613/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "CIR-AMS/BTRM_Qwen2_7b_0613", + "id": "CIR-AMS/BTRM_Qwen2_7b_0613", + "developer": "CIR-AMS", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5736 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5347 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3563 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6066 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7178 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5737 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6527 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/5dc86b72-a763-4049-a30f-06b78aa0b7f5.json b/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/5dc86b72-a763-4049-a30f-06b78aa0b7f5.json new file mode 100644 index 0000000000000000000000000000000000000000..8804bbb9203f7f9e0429bf65851782e27823874e --- /dev/null +++ b/data/reward-bench/CIR-AMS/BTRM_Qwen2_7b_0613/5dc86b72-a763-4049-a30f-06b78aa0b7f5.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/CIR-AMS_BTRM_Qwen2_7b_0613/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "CIR-AMS/BTRM_Qwen2_7b_0613", + "id": "CIR-AMS/BTRM_Qwen2_7b_0613", + "developer": "CIR-AMS", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8172 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9749 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5724 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9014 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8775 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7029 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/CohereForAI/c4ai-command-r-plus/11d9eabf-0098-48bb-8921-990c8f5a3c3e.json b/data/reward-bench/CohereForAI/c4ai-command-r-plus/11d9eabf-0098-48bb-8921-990c8f5a3c3e.json new file mode 100644 index 0000000000000000000000000000000000000000..e877753e254d53fe0f4141ddae44ba48332f8f35 --- /dev/null +++ b/data/reward-bench/CohereForAI/c4ai-command-r-plus/11d9eabf-0098-48bb-8921-990c8f5a3c3e.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/CohereForAI_c4ai-command-r-plus/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "CohereForAI/c4ai-command-r-plus", + "id": "CohereForAI/c4ai-command-r-plus", + "developer": "CohereForAI", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7057 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9511 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5757 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5986 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.704 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6924 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/HFXM/RAMO-Llama3.1-8B/08a939c2-5e4a-4563-928e-771301893f7a.json b/data/reward-bench/HFXM/RAMO-Llama3.1-8B/08a939c2-5e4a-4563-928e-771301893f7a.json new file mode 100644 index 0000000000000000000000000000000000000000..0d71e322042e7606cd85e1209170220e3013e809 --- /dev/null +++ b/data/reward-bench/HFXM/RAMO-Llama3.1-8B/08a939c2-5e4a-4563-928e-771301893f7a.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/HFXM_RAMO-Llama3.1-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "HFXM/RAMO-Llama3.1-8B", + "id": "HFXM/RAMO-Llama3.1-8B", + "developer": "HFXM", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6917 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6547 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.375 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5628 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9756 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9071 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6752 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/HuggingFaceH4/starchat2-15b-v0.1/92e00f5b-c76f-4e2b-a122-473ac1adf59b.json b/data/reward-bench/HuggingFaceH4/starchat2-15b-v0.1/92e00f5b-c76f-4e2b-a122-473ac1adf59b.json new file mode 100644 index 0000000000000000000000000000000000000000..68c2a21c6d2fe583bc77b6c39585c478cbd8d490 --- /dev/null +++ b/data/reward-bench/HuggingFaceH4/starchat2-15b-v0.1/92e00f5b-c76f-4e2b-a122-473ac1adf59b.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/HuggingFaceH4_starchat2-15b-v0.1/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "HuggingFaceH4/starchat2-15b-v0.1", + "id": "HuggingFaceH4/starchat2-15b-v0.1", + "developer": "HuggingFaceH4", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7322 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9385 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5548 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7095 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8159 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5525 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/HuggingFaceH4/zephyr-7b-alpha/f4346c54-0acd-4069-aa28-bc8d0dc86a04.json b/data/reward-bench/HuggingFaceH4/zephyr-7b-alpha/f4346c54-0acd-4069-aa28-bc8d0dc86a04.json new file mode 100644 index 0000000000000000000000000000000000000000..29bd9f96ce573e34f7bf48cc6641907ef9be831f --- /dev/null +++ b/data/reward-bench/HuggingFaceH4/zephyr-7b-alpha/f4346c54-0acd-4069-aa28-bc8d0dc86a04.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/HuggingFaceH4_zephyr-7b-alpha/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "HuggingFaceH4/zephyr-7b-alpha", + "id": "HuggingFaceH4/zephyr-7b-alpha", + "developer": "HuggingFaceH4", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7392 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9162 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.625 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7662 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7514 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5353 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/HuggingFaceH4/zephyr-7b-beta/c1f37416-e5ce-4eb1-8694-733dc5669e43.json b/data/reward-bench/HuggingFaceH4/zephyr-7b-beta/c1f37416-e5ce-4eb1-8694-733dc5669e43.json new file mode 100644 index 0000000000000000000000000000000000000000..1890321c25fc74b0623682a804fdbd4e2e948714 --- /dev/null +++ b/data/reward-bench/HuggingFaceH4/zephyr-7b-beta/c1f37416-e5ce-4eb1-8694-733dc5669e43.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/HuggingFaceH4_zephyr-7b-beta/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "HuggingFaceH4/zephyr-7b-beta", + "id": "HuggingFaceH4/zephyr-7b-beta", + "developer": "HuggingFaceH4", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7281 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9525 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6272 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6568 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7789 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5216 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/HuggingFaceH4/zephyr-7b-gemma-v0.1/bccffc93-1b2f-48ad-b3ae-81edd911e3e5.json b/data/reward-bench/HuggingFaceH4/zephyr-7b-gemma-v0.1/bccffc93-1b2f-48ad-b3ae-81edd911e3e5.json new file mode 100644 index 0000000000000000000000000000000000000000..b342cdc255c5e17daaa3fff9383aa0a42ed00571 --- /dev/null +++ b/data/reward-bench/HuggingFaceH4/zephyr-7b-gemma-v0.1/bccffc93-1b2f-48ad-b3ae-81edd911e3e5.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/HuggingFaceH4_zephyr-7b-gemma-v0.1/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "HuggingFaceH4/zephyr-7b-gemma-v0.1", + "id": "HuggingFaceH4/zephyr-7b-gemma-v0.1", + "developer": "HuggingFaceH4", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6758 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9581 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4956 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5824 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7463 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5171 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Nexusflow/Starling-RM-34B/00d734b7-cfeb-4929-aea5-d905e1d80a59.json b/data/reward-bench/Nexusflow/Starling-RM-34B/00d734b7-cfeb-4929-aea5-d905e1d80a59.json new file mode 100644 index 0000000000000000000000000000000000000000..45b4688a1650c470bb9fa971720f13810814b078 --- /dev/null +++ b/data/reward-bench/Nexusflow/Starling-RM-34B/00d734b7-cfeb-4929-aea5-d905e1d80a59.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Nexusflow_Starling-RM-34B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Nexusflow/Starling-RM-34B", + "id": "Nexusflow/Starling-RM-34B", + "developer": "Nexusflow", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8133 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9693 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5724 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.877 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8845 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7137 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Nexusflow/Starling-RM-34B/3069e3db-2485-4dec-aada-28e9d01860e7.json b/data/reward-bench/Nexusflow/Starling-RM-34B/3069e3db-2485-4dec-aada-28e9d01860e7.json new file mode 100644 index 0000000000000000000000000000000000000000..84dd16982c10e4989998c6a780af0ec097f6b20f --- /dev/null +++ b/data/reward-bench/Nexusflow/Starling-RM-34B/3069e3db-2485-4dec-aada-28e9d01860e7.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Nexusflow_Starling-RM-34B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Nexusflow/Starling-RM-34B", + "id": "Nexusflow/Starling-RM-34B", + "developer": "Nexusflow", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4553 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4589 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3187 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6175 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7556 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4808 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.1004 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-32B/8b36c417-a84d-450a-851d-cf3751dd0be6.json b/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-32B/8b36c417-a84d-450a-851d-cf3751dd0be6.json new file mode 100644 index 0000000000000000000000000000000000000000..424df44edc142d443d38ceec09b11967e23c07ab --- /dev/null +++ b/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-32B/8b36c417-a84d-450a-851d-cf3751dd0be6.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/R-I-S-E_RISE-Judge-Qwen2.5-32B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "R-I-S-E/RISE-Judge-Qwen2.5-32B", + "id": "R-I-S-E/RISE-Judge-Qwen2.5-32B", + "developer": "R-I-S-E", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9266 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9665 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8333 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9189 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9877 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-7B/b3fd4400-bcba-4ffc-9c6d-586a552f6cf4.json b/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-7B/b3fd4400-bcba-4ffc-9c6d-586a552f6cf4.json new file mode 100644 index 0000000000000000000000000000000000000000..6129331dae0748c5a70e7362ceeaf8cc303811c9 --- /dev/null +++ b/data/reward-bench/R-I-S-E/RISE-Judge-Qwen2.5-7B/b3fd4400-bcba-4ffc-9c6d-586a552f6cf4.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/R-I-S-E_RISE-Judge-Qwen2.5-7B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "R-I-S-E/RISE-Judge-Qwen2.5-7B", + "id": "R-I-S-E/RISE-Judge-Qwen2.5-7B", + "developer": "R-I-S-E", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8819 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9218 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7654 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8797 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9608 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/3fb7c433-26a7-497b-b235-8b714a316457.json b/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/3fb7c433-26a7-497b-b235-8b714a316457.json new file mode 100644 index 0000000000000000000000000000000000000000..d51deb64c07d718fd2495acc8e23cd968aaf61b3 --- /dev/null +++ b/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/3fb7c433-26a7-497b-b235-8b714a316457.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/RLHFlow_ArmoRM-Llama3-8B-v0.1/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "RLHFlow/ArmoRM-Llama3-8B-v0.1", + "id": "RLHFlow/ArmoRM-Llama3-8B-v0.1", + "developer": "RLHFlow", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.886 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9693 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7675 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9054 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9735 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7429 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/f669c7a1-4b44-40cb-8a48-9ed18b9fe45b.json b/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/f669c7a1-4b44-40cb-8a48-9ed18b9fe45b.json new file mode 100644 index 0000000000000000000000000000000000000000..7207bac583cdfb8011926673bcec788da643f290 --- /dev/null +++ b/data/reward-bench/RLHFlow/ArmoRM-Llama3-8B-v0.1/f669c7a1-4b44-40cb-8a48-9ed18b9fe45b.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/RLHFlow_ArmoRM-Llama3-8B-v0.1/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "RLHFlow/ArmoRM-Llama3-8B-v0.1", + "id": "RLHFlow/ArmoRM-Llama3-8B-v0.1", + "developer": "RLHFlow", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6646 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6568 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4188 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6612 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8222 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7657 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6629 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/RLHFlow/LLaMA3-iterative-DPO-final/d73e8f90-5cf3-4576-86af-cfa031ab7eef.json b/data/reward-bench/RLHFlow/LLaMA3-iterative-DPO-final/d73e8f90-5cf3-4576-86af-cfa031ab7eef.json new file mode 100644 index 0000000000000000000000000000000000000000..0a35dafe75ecaad832a65569dc77c65f260b00a9 --- /dev/null +++ b/data/reward-bench/RLHFlow/LLaMA3-iterative-DPO-final/d73e8f90-5cf3-4576-86af-cfa031ab7eef.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/RLHFlow_LLaMA3-iterative-DPO-final/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "RLHFlow/LLaMA3-iterative-DPO-final", + "id": "RLHFlow/LLaMA3-iterative-DPO-final", + "developer": "RLHFlow", + "additional_details": { + "model_type": "DPO" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6783 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.838 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5921 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7865 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6161 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4392 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/RLHFlow/RewardModel-Mistral-7B-for-DPA-v1/3cfd7f8f-538e-4f00-9d9f-59da5a5f3669.json b/data/reward-bench/RLHFlow/RewardModel-Mistral-7B-for-DPA-v1/3cfd7f8f-538e-4f00-9d9f-59da5a5f3669.json new file mode 100644 index 0000000000000000000000000000000000000000..ca2276ede363955b7e9976df9d81c8d9c378f866 --- /dev/null +++ b/data/reward-bench/RLHFlow/RewardModel-Mistral-7B-for-DPA-v1/3cfd7f8f-538e-4f00-9d9f-59da5a5f3669.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/RLHFlow_RewardModel-Mistral-7B-for-DPA-v1/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "RLHFlow/RewardModel-Mistral-7B-for-DPA-v1", + "id": "RLHFlow/RewardModel-Mistral-7B-for-DPA-v1", + "developer": "RLHFlow", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6633 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8799 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4978 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7068 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5971 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6068 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/RLHFlow/pair-preference-model-LLaMA3-8B/37f5dc86-1269-497d-8005-a2f79c32602e.json b/data/reward-bench/RLHFlow/pair-preference-model-LLaMA3-8B/37f5dc86-1269-497d-8005-a2f79c32602e.json new file mode 100644 index 0000000000000000000000000000000000000000..4eadc642c0661659c3f2b6fb671a44d884a97c0f --- /dev/null +++ b/data/reward-bench/RLHFlow/pair-preference-model-LLaMA3-8B/37f5dc86-1269-497d-8005-a2f79c32602e.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/RLHFlow_pair-preference-model-LLaMA3-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "RLHFlow/pair-preference-model-LLaMA3-8B", + "id": "RLHFlow/pair-preference-model-LLaMA3-8B", + "developer": "RLHFlow", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8575 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9832 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6579 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8973 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9473 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7458 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Salesforce/SFR-LLaMa-3.1-70B-Judge-r/4aac9f1c-f80b-4e5c-843a-6ad82ec2332f.json b/data/reward-bench/Salesforce/SFR-LLaMa-3.1-70B-Judge-r/4aac9f1c-f80b-4e5c-843a-6ad82ec2332f.json new file mode 100644 index 0000000000000000000000000000000000000000..d56f088dcaeeb2989b6dfbb2568f04b130445698 --- /dev/null +++ b/data/reward-bench/Salesforce/SFR-LLaMa-3.1-70B-Judge-r/4aac9f1c-f80b-4e5c-843a-6ad82ec2332f.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Salesforce_SFR-LLaMa-3.1-70B-Judge-r/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Salesforce/SFR-LLaMa-3.1-70B-Judge-r", + "id": "Salesforce/SFR-LLaMa-3.1-70B-Judge-r", + "developer": "Salesforce", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9272 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9693 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8476 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9162 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9757 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Salesforce/SFR-LLaMa-3.1-8B-Judge-r/6bc865a2-39a5-404c-aa1a-aa56c7449ff5.json b/data/reward-bench/Salesforce/SFR-LLaMa-3.1-8B-Judge-r/6bc865a2-39a5-404c-aa1a-aa56c7449ff5.json new file mode 100644 index 0000000000000000000000000000000000000000..8dde811139dfcd1cf7a87a9979ac3658b418ea73 --- /dev/null +++ b/data/reward-bench/Salesforce/SFR-LLaMa-3.1-8B-Judge-r/6bc865a2-39a5-404c-aa1a-aa56c7449ff5.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Salesforce_SFR-LLaMa-3.1-8B-Judge-r/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Salesforce/SFR-LLaMa-3.1-8B-Judge-r", + "id": "Salesforce/SFR-LLaMa-3.1-8B-Judge-r", + "developer": "Salesforce", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8865 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9553 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7774 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8622 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9513 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Salesforce/SFR-nemo-12B-Judge-r/fb05dcc0-deac-4c16-9716-f96999d74f45.json b/data/reward-bench/Salesforce/SFR-nemo-12B-Judge-r/fb05dcc0-deac-4c16-9716-f96999d74f45.json new file mode 100644 index 0000000000000000000000000000000000000000..fe19e5ee0639ff27a15950798a0993ad77164f98 --- /dev/null +++ b/data/reward-bench/Salesforce/SFR-nemo-12B-Judge-r/fb05dcc0-deac-4c16-9716-f96999d74f45.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Salesforce_SFR-nemo-12B-Judge-r/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Salesforce/SFR-nemo-12B-Judge-r", + "id": "Salesforce/SFR-nemo-12B-Judge-r", + "developer": "Salesforce", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9027 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9721 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8224 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8649 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9513 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-70B/50b016c7-2199-44e1-b553-e4c3faaf9230.json b/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-70B/50b016c7-2199-44e1-b553-e4c3faaf9230.json new file mode 100644 index 0000000000000000000000000000000000000000..600bc06415318967f120b437042b18a64fc258fc --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-70B/50b016c7-2199-44e1-b553-e4c3faaf9230.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Critic-Llama-3.1-70B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Critic-Llama-3.1-70B", + "id": "Skywork/Skywork-Critic-Llama-3.1-70B", + "developer": "Skywork", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9331 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9665 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8794 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9311 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9554 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-8B/d2f9d005-423e-450b-b712-81d8163ca142.json b/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-8B/d2f9d005-423e-450b-b712-81d8163ca142.json new file mode 100644 index 0000000000000000000000000000000000000000..218205d966af7fbf1f8f5def01aead89aab95236 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Critic-Llama-3.1-8B/d2f9d005-423e-450b-b712-81d8163ca142.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Critic-Llama-3.1-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Critic-Llama-3.1-8B", + "id": "Skywork/Skywork-Critic-Llama-3.1-8B", + "developer": "Skywork", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8896 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9358 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8136 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9108 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.898 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/911b3350-6303-4ce6-97d4-cdcb41350c6b.json b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/911b3350-6303-4ce6-97d4-cdcb41350c6b.json new file mode 100644 index 0000000000000000000000000000000000000000..2e50d30ded203831348f3cb4fed00615684e10ca --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/911b3350-6303-4ce6-97d4-cdcb41350c6b.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-Gemma-2-27B-v0.2/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Gemma-2-27B-v0.2", + "id": "Skywork/Skywork-Reward-Gemma-2-27B-v0.2", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7531 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7674 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.375 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6721 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9689 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9172 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8182 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/a23fae56-cb37-41d5-bba2-185e810bd39c.json b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/a23fae56-cb37-41d5-bba2-185e810bd39c.json new file mode 100644 index 0000000000000000000000000000000000000000..dcb11d0fdec01e0fbc3cad7f4a0914f07860b678 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B-v0.2/a23fae56-cb37-41d5-bba2-185e810bd39c.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Reward-Gemma-2-27B-v0.2/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Gemma-2-27B-v0.2", + "id": "Skywork/Skywork-Reward-Gemma-2-27B-v0.2", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9426 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9609 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8991 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9297 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9807 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/6ed31790-9f0a-472a-9ae6-915519ebba5c.json b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/6ed31790-9f0a-472a-9ae6-915519ebba5c.json new file mode 100644 index 0000000000000000000000000000000000000000..c07506d4b820c91fc736aa4235e517c602a577fa --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/6ed31790-9f0a-472a-9ae6-915519ebba5c.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Reward-Gemma-2-27B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Gemma-2-27B", + "id": "Skywork/Skywork-Reward-Gemma-2-27B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.938 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9581 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9145 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9189 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9606 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/a733a7d4-cf16-4a65-94a7-1654ebb08d36.json b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/a733a7d4-cf16-4a65-94a7-1654ebb08d36.json new file mode 100644 index 0000000000000000000000000000000000000000..fd0b3138cb2aed7f44efa47b7a99a79f2344c00d --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Gemma-2-27B/a733a7d4-cf16-4a65-94a7-1654ebb08d36.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-Gemma-2-27B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Gemma-2-27B", + "id": "Skywork/Skywork-Reward-Gemma-2-27B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7576 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7368 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4031 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7049 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9422 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9323 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8261 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/224cc1cf-8b1c-43d5-bc7d-8e2592d1c1d9.json b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/224cc1cf-8b1c-43d5-bc7d-8e2592d1c1d9.json new file mode 100644 index 0000000000000000000000000000000000000000..20fc98498f814464ff2722b60eeba65bf51bf5a1 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/224cc1cf-8b1c-43d5-bc7d-8e2592d1c1d9.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-Llama-3.1-8B-v0.2/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Llama-3.1-8B-v0.2", + "id": "Skywork/Skywork-Reward-Llama-3.1-8B-v0.2", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7175 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6968 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4062 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6011 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9422 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9414 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7169 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/88df99b5-4481-4878-8ba0-2db2abf20e2a.json b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/88df99b5-4481-4878-8ba0-2db2abf20e2a.json new file mode 100644 index 0000000000000000000000000000000000000000..5e88115382d1e81035b86296b9857edf454a89ab --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B-v0.2/88df99b5-4481-4878-8ba0-2db2abf20e2a.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Reward-Llama-3.1-8B-v0.2/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Llama-3.1-8B-v0.2", + "id": "Skywork/Skywork-Reward-Llama-3.1-8B-v0.2", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9313 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9469 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8838 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.927 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9675 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/894b5d74-829b-42e6-87f1-4695cbe14dbb.json b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/894b5d74-829b-42e6-87f1-4695cbe14dbb.json new file mode 100644 index 0000000000000000000000000000000000000000..2b449d0d0100835725ff1515f636b07d4ed10379 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/894b5d74-829b-42e6-87f1-4695cbe14dbb.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-Reward-Llama-3.1-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Llama-3.1-8B", + "id": "Skywork/Skywork-Reward-Llama-3.1-8B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9252 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9581 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8728 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9081 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.962 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/bdbb97a0-be05-4992-a23b-78c289306dd7.json b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/bdbb97a0-be05-4992-a23b-78c289306dd7.json new file mode 100644 index 0000000000000000000000000000000000000000..a19fed56599f4b18dd76eb696fbc4a66b4d56905 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-Llama-3.1-8B/bdbb97a0-be05-4992-a23b-78c289306dd7.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-Llama-3.1-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-Llama-3.1-8B", + "id": "Skywork/Skywork-Reward-Llama-3.1-8B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7314 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6989 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.425 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6284 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9333 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9616 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.741 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.1-8B/3a7e9977-82e4-4337-ba1d-98b9a800c830.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.1-8B/3a7e9977-82e4-4337-ba1d-98b9a800c830.json new file mode 100644 index 0000000000000000000000000000000000000000..6ffce761829cbc4eade23620e141af34a09b0a64 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.1-8B/3a7e9977-82e4-4337-ba1d-98b9a800c830.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Llama-3.1-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Llama-3.1-8B", + "id": "Skywork/Skywork-Reward-V2-Llama-3.1-8B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8413 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8463 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6625 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.776 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9667 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9838 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8124 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-1B/fbb40488-0265-43cb-bea0-4ea5e214385d.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-1B/fbb40488-0265-43cb-bea0-4ea5e214385d.json new file mode 100644 index 0000000000000000000000000000000000000000..675ca0b56257b295c95ca596adcd972995075b11 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-1B/fbb40488-0265-43cb-bea0-4ea5e214385d.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Llama-3.2-1B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Llama-3.2-1B", + "id": "Skywork/Skywork-Reward-V2-Llama-3.2-1B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6438 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6084 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4562 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6011 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8733 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8929 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4306 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-3B/3c8a928f-1e0c-4987-bd95-1881b34d5937.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-3B/3c8a928f-1e0c-4987-bd95-1881b34d5937.json new file mode 100644 index 0000000000000000000000000000000000000000..ae6ac1acc4685e5b7c1b2f7d37c4c98c16a3ec36 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Llama-3.2-3B/3c8a928f-1e0c-4987-bd95-1881b34d5937.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Llama-3.2-3B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Llama-3.2-3B", + "id": "Skywork/Skywork-Reward-V2-Llama-3.2-3B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7466 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7621 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4562 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.694 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9311 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9596 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6768 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-0.6B/c065f1fc-2a43-4567-b2c9-1e15b98ca56f.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-0.6B/c065f1fc-2a43-4567-b2c9-1e15b98ca56f.json new file mode 100644 index 0000000000000000000000000000000000000000..65f9d188214dcb4addd6d415dc262402fd016028 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-0.6B/c065f1fc-2a43-4567-b2c9-1e15b98ca56f.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Qwen3-0.6B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Qwen3-0.6B", + "id": "Skywork/Skywork-Reward-V2-Qwen3-0.6B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6125 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.58 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7158 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8444 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7949 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3397 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-1.7B/0b66d928-c8fc-4972-8bb6-afc54f32d096.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-1.7B/0b66d928-c8fc-4972-8bb6-afc54f32d096.json new file mode 100644 index 0000000000000000000000000000000000000000..af67a0d33054d8c884f27608406bc204092e4d5b --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-1.7B/0b66d928-c8fc-4972-8bb6-afc54f32d096.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Qwen3-1.7B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Qwen3-1.7B", + "id": "Skywork/Skywork-Reward-V2-Qwen3-1.7B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6818 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6568 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4437 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7268 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8911 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8848 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4872 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-4B/1fa8b2de-0d05-406f-8ce9-369a1f168137.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-4B/1fa8b2de-0d05-406f-8ce9-369a1f168137.json new file mode 100644 index 0000000000000000000000000000000000000000..249b4fc385fba351d755c157b625cca1a2f406d0 --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-4B/1fa8b2de-0d05-406f-8ce9-369a1f168137.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Qwen3-4B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Qwen3-4B", + "id": "Skywork/Skywork-Reward-V2-Qwen3-4B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7551 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7737 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4625 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7322 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9222 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9657 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6743 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-8B/46d17060-eb14-43e5-a75e-d1599c05c63a.json b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-8B/46d17060-eb14-43e5-a75e-d1599c05c63a.json new file mode 100644 index 0000000000000000000000000000000000000000..376b4c595e476654e4b8fee1a47a3e7b9d2a0b9c --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-Reward-V2-Qwen3-8B/46d17060-eb14-43e5-a75e-d1599c05c63a.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-Reward-V2-Qwen3-8B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-Reward-V2-Qwen3-8B", + "id": "Skywork/Skywork-Reward-V2-Qwen3-8B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7837 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7989 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7705 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.94 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9636 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7294 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-VL-Reward-7B/4d5de0f1-6790-499f-8701-d230d33f6cd2.json b/data/reward-bench/Skywork/Skywork-VL-Reward-7B/4d5de0f1-6790-499f-8701-d230d33f6cd2.json new file mode 100644 index 0000000000000000000000000000000000000000..6935699c691da1a4a8bb40c853a887264c2c1afc --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-VL-Reward-7B/4d5de0f1-6790-499f-8701-d230d33f6cd2.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/Skywork_Skywork-VL-Reward-7B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-VL-Reward-7B", + "id": "Skywork/Skywork-VL-Reward-7B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9007 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8994 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.875 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9108 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9176 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/Skywork/Skywork-VL-Reward-7B/cb2bec6d-2413-4a55-a3fe-c44ac2ae3756.json b/data/reward-bench/Skywork/Skywork-VL-Reward-7B/cb2bec6d-2413-4a55-a3fe-c44ac2ae3756.json new file mode 100644 index 0000000000000000000000000000000000000000..455c143690bcdf1f0b8d67f765db9ad7d614c4db --- /dev/null +++ b/data/reward-bench/Skywork/Skywork-VL-Reward-7B/cb2bec6d-2413-4a55-a3fe-c44ac2ae3756.json @@ -0,0 +1,152 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench-2/Skywork_Skywork-VL-Reward-7B/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench 2", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "Skywork/Skywork-VL-Reward-7B", + "id": "Skywork/Skywork-VL-Reward-7B", + "developer": "Skywork", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench 2 Score (mean of all metrics)", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6885 + } + }, + { + "evaluation_name": "Factuality", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Factuality score - measures factual accuracy", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6063 + } + }, + { + "evaluation_name": "Precise IF", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Precise Instruction Following score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.35 + } + }, + { + "evaluation_name": "Math", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Math score - measures mathematical reasoning", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6339 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Safety score - measures safety awareness", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8911 + } + }, + { + "evaluation_name": "Focus", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Focus score - measures response focus", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8909 + } + }, + { + "evaluation_name": "Ties", + "source_data": { + "dataset_name": "RewardBench 2", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench-2-results" + }, + "metric_config": { + "evaluation_description": "Ties score - ability to identify tie cases", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7586 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/internlm/internlm2-7b-reward/39bf79ae-fac1-4384-9694-7a8302cf985d.json b/data/reward-bench/internlm/internlm2-7b-reward/39bf79ae-fac1-4384-9694-7a8302cf985d.json new file mode 100644 index 0000000000000000000000000000000000000000..e26f1ebcb9af9d0ed1eb59393884cc91b549dcce --- /dev/null +++ b/data/reward-bench/internlm/internlm2-7b-reward/39bf79ae-fac1-4384-9694-7a8302cf985d.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/internlm_internlm2-7b-reward/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "internlm/internlm2-7b-reward", + "id": "internlm/internlm2-7b-reward", + "developer": "internlm", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8759 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9916 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6952 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8716 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9453 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/llm-blender/PairRM-hf/1bd4b100-11ac-4810-8763-b58cced3c51a.json b/data/reward-bench/llm-blender/PairRM-hf/1bd4b100-11ac-4810-8763-b58cced3c51a.json new file mode 100644 index 0000000000000000000000000000000000000000..440d1f3770aaf53814f44021442c64f7ba3b8c1f --- /dev/null +++ b/data/reward-bench/llm-blender/PairRM-hf/1bd4b100-11ac-4810-8763-b58cced3c51a.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/llm-blender_PairRM-hf/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "llm-blender/PairRM-hf", + "id": "llm-blender/PairRM-hf", + "developer": "llm-blender", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6087 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9022 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5219 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.477 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4898 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6961 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3-70B-Instruct/a7610260-32fc-46af-bc03-341803667ae4.json b/data/reward-bench/meta-llama/Meta-Llama-3-70B-Instruct/a7610260-32fc-46af-bc03-341803667ae4.json new file mode 100644 index 0000000000000000000000000000000000000000..4b19f5bb699023f278bb798ffe0c1a248b947950 --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3-70B-Instruct/a7610260-32fc-46af-bc03-341803667ae4.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3-70B-Instruct/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3-70B-Instruct", + "id": "meta-llama/Meta-Llama-3-70B-Instruct", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7627 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9763 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5888 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7297 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7854 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7035 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3-8B-Instruct/1eedc453-4dc1-4dcd-bb0b-99b8857f7f3a.json b/data/reward-bench/meta-llama/Meta-Llama-3-8B-Instruct/1eedc453-4dc1-4dcd-bb0b-99b8857f7f3a.json new file mode 100644 index 0000000000000000000000000000000000000000..6a6cb2802002b595468dfcf5a11da4e416ad0582 --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3-8B-Instruct/1eedc453-4dc1-4dcd-bb0b-99b8857f7f3a.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3-8B-Instruct/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3-8B-Instruct", + "id": "meta-llama/Meta-Llama-3-8B-Instruct", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.645 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8547 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4156 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6797 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6482 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6082 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo/8eba2226-b457-433b-aeea-e341a7fa3311.json b/data/reward-bench/meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo/8eba2226-b457-433b-aeea-e341a7fa3311.json new file mode 100644 index 0000000000000000000000000000000000000000..a72a79355a6300f3a32d98f3e6e2a6a0f2e619e5 --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo/8eba2226-b457-433b-aeea-e341a7fa3311.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3.1-405B-Instruct-Turbo/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", + "id": "meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8412 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9721 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7456 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7757 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8715 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo/a718af93-91ef-455e-be69-aa8072719f77.json b/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo/a718af93-91ef-455e-be69-aa8072719f77.json new file mode 100644 index 0000000000000000000000000000000000000000..e90e4e734c0c6d136053bb1fe604987e2be9d4d5 --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo/a718af93-91ef-455e-be69-aa8072719f77.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3.1-70B-Instruct-Turbo/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo", + "id": "meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7808 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8757 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6689 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7507 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.828 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct/7134c50b-b4b7-48c7-8b70-12d2c738a328.json b/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct/7134c50b-b4b7-48c7-8b70-12d2c738a328.json new file mode 100644 index 0000000000000000000000000000000000000000..93482898a4c664645d079ac21c676c79f2ce9fa2 --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3.1-70B-Instruct/7134c50b-b4b7-48c7-8b70-12d2c738a328.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3.1-70B-Instruct/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3.1-70B-Instruct", + "id": "meta-llama/Meta-Llama-3.1-70B-Instruct", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8405 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.9721 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.7018 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8284 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8599 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo/bb5df85c-26c0-4493-91f3-86444a9d4a71.json b/data/reward-bench/meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo/bb5df85c-26c0-4493-91f3-86444a9d4a71.json new file mode 100644 index 0000000000000000000000000000000000000000..4d4ab8e02d667bfa8354c7c2b65d41f8a18b3a2e --- /dev/null +++ b/data/reward-bench/meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo/bb5df85c-26c0-4493-91f3-86444a9d4a71.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/meta-llama_Meta-Llama-3.1-8B-Instruct-Turbo/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo", + "id": "meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo", + "developer": "meta-llama", + "additional_details": { + "model_type": "Generative" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6565 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8073 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4978 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6399 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6811 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/my_model/e4b366cc-66e9-49e4-8dd1-5a64f8ec728f.json b/data/reward-bench/my_model/e4b366cc-66e9-49e4-8dd1-5a64f8ec728f.json new file mode 100644 index 0000000000000000000000000000000000000000..97205c31492b5c779acc41e9ad655a1037d65d15 --- /dev/null +++ b/data/reward-bench/my_model/e4b366cc-66e9-49e4-8dd1-5a64f8ec728f.json @@ -0,0 +1,116 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/my_model_/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "my_model/", + "id": "my_model/", + "developer": "my_model", + "additional_details": { + "model_type": "Seq. Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5267 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4553 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5592 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4392 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6532 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-large/f5186b7e-7874-4695-9eee-cbf8c399a2f3.json b/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-large/f5186b7e-7874-4695-9eee-cbf8c399a2f3.json new file mode 100644 index 0000000000000000000000000000000000000000..f1930c846015072ce597d21e0315d0a96602d392 --- /dev/null +++ b/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-large/f5186b7e-7874-4695-9eee-cbf8c399a2f3.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/stanfordnlp_SteamSHP-flan-t5-large/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "stanfordnlp/SteamSHP-flan-t5-large", + "id": "stanfordnlp/SteamSHP-flan-t5-large", + "developer": "stanfordnlp", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.4962 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8575 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3311 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3743 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3563 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6273 + } + } + ] +} \ No newline at end of file diff --git a/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-xl/52c878a9-a1a3-4491-8599-7e8ada1c4c4f.json b/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-xl/52c878a9-a1a3-4491-8599-7e8ada1c4c4f.json new file mode 100644 index 0000000000000000000000000000000000000000..534ef09f040e3d1b634fa6036ed5f0cf613b2f9f --- /dev/null +++ b/data/reward-bench/stanfordnlp/SteamSHP-flan-t5-xl/52c878a9-a1a3-4491-8599-7e8ada1c4c4f.json @@ -0,0 +1,134 @@ +{ + "schema_version": "0.2.1", + "evaluation_id": "reward-bench/stanfordnlp_SteamSHP-flan-t5-xl/1782803530.6795423", + "retrieved_timestamp": "1782803530.6795423", + "source_metadata": { + "source_name": "RewardBench", + "source_type": "documentation", + "source_organization_name": "Allen Institute for AI", + "source_organization_url": "https://allenai.org", + "evaluator_relationship": "third_party" + }, + "eval_library": { + "name": "unknown", + "version": "unknown" + }, + "model_info": { + "name": "stanfordnlp/SteamSHP-flan-t5-xl", + "id": "stanfordnlp/SteamSHP-flan-t5-xl", + "developer": "stanfordnlp", + "additional_details": { + "model_type": "Custom Classifier" + } + }, + "evaluation_results": [ + { + "evaluation_name": "Score", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Overall RewardBench Score", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.5135 + } + }, + { + "evaluation_name": "Chat", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat accuracy - includes easy chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.8547 + } + }, + { + "evaluation_name": "Chat Hard", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Chat Hard accuracy - includes hard chat subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3684 + } + }, + { + "evaluation_name": "Safety", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Safety accuracy - includes safety subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3784 + } + }, + { + "evaluation_name": "Reasoning", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Reasoning accuracy - includes code and math subsets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.3841 + } + }, + { + "evaluation_name": "Prior Sets (0.5 weight)", + "source_data": { + "dataset_name": "RewardBench", + "source_type": "hf_dataset", + "hf_repo": "allenai/reward-bench" + }, + "metric_config": { + "evaluation_description": "Prior Sets score (weighted 0.5) - includes test sets", + "lower_is_better": false, + "score_type": "continuous", + "min_score": 0.0, + "max_score": 1.0 + }, + "score_details": { + "score": 0.6498 + } + } + ] +} \ No newline at end of file