{ "adapter_file": "adapters/primary/primary/adapter_model.safetensors", "adapter_sha256": "f0a1e400ab37136131996109ff3c041c8a8e4425e11250c5d82eee098b43baff", "adapter_trainable_parameters": 1298432, "artifact_hashes": { "adapters/primary/README.md": "b997674c5241dc4f82e15cb8ddd7c3ffef8b8bab3d31a648f3bd70bb8d5eac35", "adapters/primary/primary/adapter_config.json": "3869910184dc4955a19a1c7ca753cb0018fe9b8d5c4a49b373eec56d53f10004", "adapters/primary/primary/adapter_model.safetensors": "f0a1e400ab37136131996109ff3c041c8a8e4425e11250c5d82eee098b43baff", "checkpoint-metrics.jsonl": "1dc37f0670a6550e8daa933e9be762b4eb7cc9863bba62dcadadd452eaa5fd5d", "checkpoints/step-100/README.md": "b997674c5241dc4f82e15cb8ddd7c3ffef8b8bab3d31a648f3bd70bb8d5eac35", "checkpoints/step-100/primary/adapter_config.json": "3869910184dc4955a19a1c7ca753cb0018fe9b8d5c4a49b373eec56d53f10004", "checkpoints/step-100/primary/adapter_model.safetensors": "1d8d7a813be64eae23f083648b24ed18df1813fcca09342641c1fd9549f1749a", "checkpoints/step-200/README.md": "b997674c5241dc4f82e15cb8ddd7c3ffef8b8bab3d31a648f3bd70bb8d5eac35", "checkpoints/step-200/primary/adapter_config.json": "3869910184dc4955a19a1c7ca753cb0018fe9b8d5c4a49b373eec56d53f10004", "checkpoints/step-200/primary/adapter_model.safetensors": "694e7f0120625822365ea58e6cb2c1808579fd4a4e0977b676199ba996c34588", "checkpoints/step-50/README.md": "b997674c5241dc4f82e15cb8ddd7c3ffef8b8bab3d31a648f3bd70bb8d5eac35", "checkpoints/step-50/primary/adapter_config.json": "3869910184dc4955a19a1c7ca753cb0018fe9b8d5c4a49b373eec56d53f10004", "checkpoints/step-50/primary/adapter_model.safetensors": "f0a1e400ab37136131996109ff3c041c8a8e4425e11250c5d82eee098b43baff", "grammar-preflight.json": "bdaf72ff6d0c21ae5dcc8d772ae505c6f82bf28e32290498b6295d5ff9dff8b0", "result.json": "fc8d74ccc1cee13421f1f523d4341a02b66d05ca8da498102f70656cae137427", "training-manifest.json": "d07de5affabfe47c22cd7e99ef9982829027f4ea24b6bab8748981f6c605ab94" }, "artifact_root": "/jump-runs/gemma-primary-exploratory/d07de5affabfe47c22cd7e99ef9982829027f4ea24b6bab8748981f6c605ab94", "base_trainable_parameters": 0, "checkpoints": [ { "adapter_dir": "checkpoints/step-50/primary", "adapter_sha256": "f0a1e400ab37136131996109ff3c041c8a8e4425e11250c5d82eee098b43baff", "answer_token_accuracy": 0.9217171717171717, "decoded": [ { "episode_id": "60134c6e5af0ad59742d", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,0,0]}" }, { "episode_id": "d614ab65659bd62b339e", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,0,1,1,1]}" }, { "episode_id": "1781a89d8e6cd33650a3", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "ccea8fb5cc95448fdfac", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,0]}" }, { "episode_id": "1ef1104d156c78f54690", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,0]}" }, { "episode_id": "94ca013f0357f8526974", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,0,1,1]}" }, { "episode_id": "6ecd6b981be1a15eaa46", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "655b63a4c832454e56f6", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "d3cdbeb5336553a8f592", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,1,0,0]}" }, { "episode_id": "92fe86aaca92265e076f", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,0,1,1,0]}" }, { "episode_id": "6d516a089a32fbbbea1b", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,0,0,0,1]}" }, { "episode_id": "4573813980de6888fc9d", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,1,0,1,0]}" } ], "examples": 12, "generated_exact_json_accuracy": 0.3333333333333333, "generated_json_parse_rate": 1.0, "step": 50, "teacher_forced_loss": 0.31487923860549927 }, { "adapter_dir": "checkpoints/step-100/primary", "adapter_sha256": "1d8d7a813be64eae23f083648b24ed18df1813fcca09342641c1fd9549f1749a", "answer_token_accuracy": 0.9621212121212122, "decoded": [ { "episode_id": "60134c6e5af0ad59742d", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,0,1,0,0]}" }, { "episode_id": "d614ab65659bd62b339e", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,0,1,1,1]}" }, { "episode_id": "1781a89d8e6cd33650a3", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "ccea8fb5cc95448fdfac", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,0]}" }, { "episode_id": "1ef1104d156c78f54690", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "94ca013f0357f8526974", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,0,0,1,1]}" }, { "episode_id": "6ecd6b981be1a15eaa46", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "655b63a4c832454e56f6", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "d3cdbeb5336553a8f592", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,1,0,0]}" }, { "episode_id": "92fe86aaca92265e076f", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,0,0,1,1,0]}" }, { "episode_id": "6d516a089a32fbbbea1b", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,0,0,0,1]}" }, { "episode_id": "4573813980de6888fc9d", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,1,0,1,0]}" } ], "examples": 12, "generated_exact_json_accuracy": 0.25, "generated_json_parse_rate": 1.0, "step": 100, "teacher_forced_loss": 0.05907239951193333 }, { "adapter_dir": "checkpoints/step-200/primary", "adapter_sha256": "694e7f0120625822365ea58e6cb2c1808579fd4a4e0977b676199ba996c34588", "answer_token_accuracy": 0.9671717171717171, "decoded": [ { "episode_id": "60134c6e5af0ad59742d", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,0,0]}" }, { "episode_id": "d614ab65659bd62b339e", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,0,0,1,1,1]}" }, { "episode_id": "1781a89d8e6cd33650a3", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "ccea8fb5cc95448fdfac", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,0]}" }, { "episode_id": "1ef1104d156c78f54690", "exact": true, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "94ca013f0357f8526974", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,0,1,1]}" }, { "episode_id": "6ecd6b981be1a15eaa46", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,1,0,0,1]}" }, { "episode_id": "655b63a4c832454e56f6", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "d3cdbeb5336553a8f592", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,1,1,0,0]}" }, { "episode_id": "92fe86aaca92265e076f", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,0,1,1,0]}" }, { "episode_id": "6d516a089a32fbbbea1b", "exact": false, "text": "{\"law\":{\"different\":\"attract\",\"exponent\":2,\"same\":\"repel\"},\"partition\":[0,1,0,0,0,1]}" }, { "episode_id": "4573813980de6888fc9d", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,0,1,0]}" } ], "examples": 12, "generated_exact_json_accuracy": 0.08333333333333333, "generated_json_parse_rate": 1.0, "step": 200, "teacher_forced_loss": 0.06225178204476833 } ], "claim_label": "exploratory product recovery; not confirmatory, G3, or mechanistic evidence", "dataset_sha256": "10b3d9cad161d415914e62795136675339a29c3cb141a4169b413da3dad75802", "demo_usable": true, "duration_seconds": 407.859539173, "environment": { "cuda": "13.0", "gpu": "NVIDIA H100 80GB HBM3", "package_lock": { "accelerate": "1.14.0", "huggingface_hub": "1.14.0", "peft": "0.20.0", "pillow": "12.3.0", "python": "3.12", "safetensors": "0.8.0", "torch": "2.11.0", "torchvision": "0.26.0", "transformers": "git+https://github.com/huggingface/transformers.git@918dbf131d0df5b46e3f6e1d96174d62aa4d16d6" }, "python": "3.12.10", "torch": "2.11.0+cu130" }, "estimated_cost_usd": 0.4474219144727809, "final_test": { "answer_token_accuracy": 0.9090909090909091, "decoded": [ { "episode_id": "ae34a0824f633da593fd", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,1]}" }, { "episode_id": "7a002ceb4dad1056e65a", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" }, { "episode_id": "9246e9536e368abc52b7", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,0,1,0,0]}" }, { "episode_id": "4955b2e1bd5fb2e31a27", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,0,1,0]}" }, { "episode_id": "a7d37707e18bc3ea163c", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,1,1,1]}" }, { "episode_id": "495e93030025b0a0eb31", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,0]}" }, { "episode_id": "e74b713b01d6d6df2781", "exact": true, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,1,0,1,1]}" }, { "episode_id": "2951145c59e4ec3d5620", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,1,0,1,1,1]}" }, { "episode_id": "5896a624cc99c4669ad3", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,0,1,0,1]}" }, { "episode_id": "820568098820a9052af4", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":2,\"same\":\"attract\"},\"partition\":[0,1,0,0,1,0]}" }, { "episode_id": "83370c8d6e932dc0fcaa", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,0,0]}" }, { "episode_id": "efef45593582c8311f4d", "exact": false, "text": "{\"law\":{\"different\":\"repel\",\"exponent\":1,\"same\":\"attract\"},\"partition\":[0,0,1,0,1,1]}" } ], "examples": 12, "generated_exact_json_accuracy": 0.25, "generated_json_parse_rate": 1.0, "teacher_forced_loss": 0.31303876638412476 }, "forecast_cost_usd": 7.8984, "grammar_sha256": "c5c67dca66f7f97fadb010a98b0c946960a4b7e96c8c08bff4ee7b2a44b8d20c", "manifest_sha256": "d07de5affabfe47c22cd7e99ef9982829027f4ea24b6bab8748981f6c605ab94", "mechanistic_evidence": false, "peak_cuda_memory_bytes": 44658674176, "schema_version": "jump.track-h-gemma-product-recovery-result/v1", "selected_step": 50, "status": "completed", "stopped_on_parse_regression": false, "tokenized_dataset_sha256": "1bfe1a46d19c8397e58f9fb8b4553524a0dce3764659966d1a1ec3c63b5d83ae" }