Hugging Face's logo

Datasets:
deeplumiere
/
EEE_datastore

Modalities:
Text
Formats:
parquet
Size:
1K - 10K
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
14

Fixed Submission for EvalBot Warning PR144

#12
by deeplumiere - opened
base: refs/heads/main
←
from: refs/pr/12
Discussion Files changed
+606
-0
Files changed (3) hide show
  1. data/vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/65f449e7-595b-4031-8364-2b24d2d6ff95.json +202 -0
  2. data/vectara-hallucination-leaderboard/microsoft/Phi-4/93cd2b22-2bd6-438a-95a6-7c32134638c5.json +202 -0
  3. data/vectara-hallucination-leaderboard/qwen/qwen3-8b/644ea67b-89da-4310-b05c-bb4064abc2ad.json +202 -0
data/vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/65f449e7-595b-4031-8364-2b24d2d6ff95.json ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/1779880742.736885",
4
+ "retrieved_timestamp": "1779880742.736885",
5
+ "evaluation_timestamp": "2025-12-10 14:57:20.585062",
6
+ "source_metadata": {
7
+ "source_name": "Vectara Hallucination Leaderboard",
8
+ "source_type": "documentation",
9
+ "source_organization_name": "Vectara",
10
+ "source_organization_url": "https://vectara.com",
11
+ "evaluator_relationship": "third_party",
12
+ "additional_details": {
13
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
14
+ "structured_results_hf_repo": "vectara/results",
15
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
16
+ "source_file": "google/gemini-2.5-flash-lite/results_2025-12-10 14:57:20.585062.json",
17
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/google/gemini-2.5-flash-lite/results_2025-12-10%2014%3A57%3A20.585062.json",
18
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
19
+ "underlying_evaluation_dataset": "Vectara Hallucination Leaderboard private evaluation dataset",
20
+ "underlying_evaluation_dataset_availability": "Private/proprietary dataset; not publicly released to avoid overfitting.",
21
+ "scoring_model": "Vectara HHEM-2.3",
22
+ "generation_temperature": "0 unless unavailable, per source documentation",
23
+ "evaluation_timestamp_source": "Parsed from the source result filename."
24
+ }
25
+ },
26
+ "model_info": {
27
+ "name": "gemini-2.5-flash-lite",
28
+ "id": "google/gemini-2.5-flash-lite",
29
+ "developer": "google",
30
+ "inference_platform": "vertex_ai",
31
+ "additional_details": {
32
+ "api_model_reference": "gemini-2.5-flash-lite",
33
+ "source_model_name": "google/gemini-2.5-flash-lite-",
34
+ "model_dtype": "float16",
35
+ "model_sha": "main",
36
+ "model_size": "large",
37
+ "accessibility": "commercial",
38
+ "deployment_type": "api"
39
+ }
40
+ },
41
+ "eval_library": {
42
+ "name": "unknown",
43
+ "version": "unknown",
44
+ "additional_details": {
45
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
46
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
47
+ "scoring_model": "Vectara HHEM-2.3"
48
+ }
49
+ },
50
+ "evaluation_results": [
51
+ {
52
+ "evaluation_result_id": "vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/1779880742.736885#hallucination_rate",
53
+ "evaluation_name": "Vectara Hallucination Leaderboard",
54
+ "source_data": {
55
+ "dataset_name": "vectara/hallucination-leaderboard",
56
+ "source_type": "url",
57
+ "url": [
58
+ "https://github.com/vectara/hallucination-leaderboard"
59
+ ]
60
+ },
61
+ "evaluation_timestamp": "2025-12-10 14:57:20.585062",
62
+ "metric_config": {
63
+ "evaluation_description": "Percentage of generated summaries judged to contain factual inconsistencies or unsupported claims.",
64
+ "metric_id": "vectara-hallucination-leaderboard.hallucination_rate",
65
+ "metric_name": "Hallucination Rate",
66
+ "metric_kind": "rate",
67
+ "metric_unit": "percent",
68
+ "metric_parameters": {},
69
+ "lower_is_better": true,
70
+ "score_type": "continuous",
71
+ "min_score": 0.0,
72
+ "max_score": 100.0,
73
+ "additional_details": {
74
+ "source_metric_key": "hallucination_rate",
75
+ "source_file": "google/gemini-2.5-flash-lite/results_2025-12-10 14:57:20.585062.json",
76
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
77
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/google/gemini-2.5-flash-lite/results_2025-12-10%2014%3A57%3A20.585062.json",
78
+ "evaluation_timestamp_source": "Parsed from the source result filename."
79
+ }
80
+ },
81
+ "score_details": {
82
+ "score": 3.3,
83
+ "details": {
84
+ "source_value_unit": "percent"
85
+ }
86
+ }
87
+ },
88
+ {
89
+ "evaluation_result_id": "vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/1779880742.736885#factual_consistency_rate",
90
+ "evaluation_name": "Vectara Hallucination Leaderboard",
91
+ "source_data": {
92
+ "dataset_name": "vectara/hallucination-leaderboard",
93
+ "source_type": "url",
94
+ "url": [
95
+ "https://github.com/vectara/hallucination-leaderboard"
96
+ ]
97
+ },
98
+ "evaluation_timestamp": "2025-12-10 14:57:20.585062",
99
+ "metric_config": {
100
+ "evaluation_description": "Percentage of generated summaries judged factually consistent.",
101
+ "metric_id": "vectara-hallucination-leaderboard.factual_consistency_rate",
102
+ "metric_name": "Factual Consistency Rate",
103
+ "metric_kind": "rate",
104
+ "metric_unit": "percent",
105
+ "metric_parameters": {},
106
+ "lower_is_better": false,
107
+ "score_type": "continuous",
108
+ "min_score": 0.0,
109
+ "max_score": 100.0,
110
+ "additional_details": {
111
+ "source_metric_key": "factual_consistency_rate",
112
+ "source_file": "google/gemini-2.5-flash-lite/results_2025-12-10 14:57:20.585062.json",
113
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
114
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/google/gemini-2.5-flash-lite/results_2025-12-10%2014%3A57%3A20.585062.json",
115
+ "evaluation_timestamp_source": "Parsed from the source result filename."
116
+ }
117
+ },
118
+ "score_details": {
119
+ "score": 96.7,
120
+ "details": {
121
+ "source_value_unit": "percent"
122
+ }
123
+ }
124
+ },
125
+ {
126
+ "evaluation_result_id": "vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/1779880742.736885#answer_rate",
127
+ "evaluation_name": "Vectara Hallucination Leaderboard",
128
+ "source_data": {
129
+ "dataset_name": "vectara/hallucination-leaderboard",
130
+ "source_type": "url",
131
+ "url": [
132
+ "https://github.com/vectara/hallucination-leaderboard"
133
+ ]
134
+ },
135
+ "evaluation_timestamp": "2025-12-10 14:57:20.585062",
136
+ "metric_config": {
137
+ "evaluation_description": "Percentage of prompts for which the model produced an answer.",
138
+ "metric_id": "vectara-hallucination-leaderboard.answer_rate",
139
+ "metric_name": "Answer Rate",
140
+ "metric_kind": "rate",
141
+ "metric_unit": "percent",
142
+ "metric_parameters": {},
143
+ "lower_is_better": false,
144
+ "score_type": "continuous",
145
+ "min_score": 0.0,
146
+ "max_score": 100.0,
147
+ "additional_details": {
148
+ "source_metric_key": "answer_rate",
149
+ "source_file": "google/gemini-2.5-flash-lite/results_2025-12-10 14:57:20.585062.json",
150
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
151
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/google/gemini-2.5-flash-lite/results_2025-12-10%2014%3A57%3A20.585062.json",
152
+ "evaluation_timestamp_source": "Parsed from the source result filename."
153
+ }
154
+ },
155
+ "score_details": {
156
+ "score": 99.5,
157
+ "details": {
158
+ "source_value_unit": "percent"
159
+ }
160
+ }
161
+ },
162
+ {
163
+ "evaluation_result_id": "vectara-hallucination-leaderboard/google/gemini-2.5-flash-lite/1779880742.736885#average_summary_length",
164
+ "evaluation_name": "Vectara Hallucination Leaderboard",
165
+ "source_data": {
166
+ "dataset_name": "vectara/hallucination-leaderboard",
167
+ "source_type": "url",
168
+ "url": [
169
+ "https://github.com/vectara/hallucination-leaderboard"
170
+ ]
171
+ },
172
+ "evaluation_timestamp": "2025-12-10 14:57:20.585062",
173
+ "metric_config": {
174
+ "evaluation_description": "Mean generated summary length in words; reported as a diagnostic metric.",
175
+ "metric_id": "vectara-hallucination-leaderboard.average_summary_length",
176
+ "metric_name": "Average Summary Length",
177
+ "metric_kind": "length",
178
+ "metric_unit": "words",
179
+ "metric_parameters": {},
180
+ "lower_is_better": false,
181
+ "score_type": "continuous",
182
+ "min_score": 0.0,
183
+ "max_score": 1000.0,
184
+ "additional_details": {
185
+ "source_metric_key": "average_summary_length",
186
+ "source_file": "google/gemini-2.5-flash-lite/results_2025-12-10 14:57:20.585062.json",
187
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
188
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/google/gemini-2.5-flash-lite/results_2025-12-10%2014%3A57%3A20.585062.json",
189
+ "evaluation_timestamp_source": "Parsed from the source result filename.",
190
+ "diagnostic_metric": "true",
191
+ "lower_is_better_note": "This is descriptive context rather than an optimization target."
192
+ }
193
+ },
194
+ "score_details": {
195
+ "score": 95.7,
196
+ "details": {
197
+ "source_value_unit": "words"
198
+ }
199
+ }
200
+ }
201
+ ]
202
+ }
data/vectara-hallucination-leaderboard/microsoft/Phi-4/93cd2b22-2bd6-438a-95a6-7c32134638c5.json ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vectara-hallucination-leaderboard/microsoft/Phi-4/1779880742.736885",
4
+ "retrieved_timestamp": "1779880742.736885",
5
+ "evaluation_timestamp": "2025-12-10 14:57:16.944171",
6
+ "source_metadata": {
7
+ "source_name": "Vectara Hallucination Leaderboard",
8
+ "source_type": "documentation",
9
+ "source_organization_name": "Vectara",
10
+ "source_organization_url": "https://vectara.com",
11
+ "evaluator_relationship": "third_party",
12
+ "additional_details": {
13
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
14
+ "structured_results_hf_repo": "vectara/results",
15
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
16
+ "source_file": "microsoft/Phi-4/results_2025-12-10 14:57:16.944171.json",
17
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/microsoft/Phi-4/results_2025-12-10%2014%3A57%3A16.944171.json",
18
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
19
+ "underlying_evaluation_dataset": "Vectara Hallucination Leaderboard private evaluation dataset",
20
+ "underlying_evaluation_dataset_availability": "Private/proprietary dataset; not publicly released to avoid overfitting.",
21
+ "scoring_model": "Vectara HHEM-2.3",
22
+ "generation_temperature": "0 unless unavailable, per source documentation",
23
+ "evaluation_timestamp_source": "Parsed from the source result filename."
24
+ }
25
+ },
26
+ "model_info": {
27
+ "name": "Phi-4",
28
+ "id": "microsoft/Phi-4",
29
+ "developer": "microsoft",
30
+ "inference_platform": "azure",
31
+ "additional_details": {
32
+ "api_model_reference": "Phi-4",
33
+ "source_model_name": "microsoft/Phi-4-",
34
+ "model_dtype": "float16",
35
+ "model_sha": "main",
36
+ "model_size": "small",
37
+ "accessibility": "open",
38
+ "deployment_type": "api"
39
+ }
40
+ },
41
+ "eval_library": {
42
+ "name": "unknown",
43
+ "version": "unknown",
44
+ "additional_details": {
45
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
46
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
47
+ "scoring_model": "Vectara HHEM-2.3"
48
+ }
49
+ },
50
+ "evaluation_results": [
51
+ {
52
+ "evaluation_result_id": "vectara-hallucination-leaderboard/microsoft/Phi-4/1779880742.736885#hallucination_rate",
53
+ "evaluation_name": "Vectara Hallucination Leaderboard",
54
+ "source_data": {
55
+ "dataset_name": "vectara/hallucination-leaderboard",
56
+ "source_type": "url",
57
+ "url": [
58
+ "https://github.com/vectara/hallucination-leaderboard"
59
+ ]
60
+ },
61
+ "evaluation_timestamp": "2025-12-10 14:57:16.944171",
62
+ "metric_config": {
63
+ "evaluation_description": "Percentage of generated summaries judged to contain factual inconsistencies or unsupported claims.",
64
+ "metric_id": "vectara-hallucination-leaderboard.hallucination_rate",
65
+ "metric_name": "Hallucination Rate",
66
+ "metric_kind": "rate",
67
+ "metric_unit": "percent",
68
+ "metric_parameters": {},
69
+ "lower_is_better": true,
70
+ "score_type": "continuous",
71
+ "min_score": 0.0,
72
+ "max_score": 100.0,
73
+ "additional_details": {
74
+ "source_metric_key": "hallucination_rate",
75
+ "source_file": "microsoft/Phi-4/results_2025-12-10 14:57:16.944171.json",
76
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
77
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/microsoft/Phi-4/results_2025-12-10%2014%3A57%3A16.944171.json",
78
+ "evaluation_timestamp_source": "Parsed from the source result filename."
79
+ }
80
+ },
81
+ "score_details": {
82
+ "score": 3.7,
83
+ "details": {
84
+ "source_value_unit": "percent"
85
+ }
86
+ }
87
+ },
88
+ {
89
+ "evaluation_result_id": "vectara-hallucination-leaderboard/microsoft/Phi-4/1779880742.736885#factual_consistency_rate",
90
+ "evaluation_name": "Vectara Hallucination Leaderboard",
91
+ "source_data": {
92
+ "dataset_name": "vectara/hallucination-leaderboard",
93
+ "source_type": "url",
94
+ "url": [
95
+ "https://github.com/vectara/hallucination-leaderboard"
96
+ ]
97
+ },
98
+ "evaluation_timestamp": "2025-12-10 14:57:16.944171",
99
+ "metric_config": {
100
+ "evaluation_description": "Percentage of generated summaries judged factually consistent.",
101
+ "metric_id": "vectara-hallucination-leaderboard.factual_consistency_rate",
102
+ "metric_name": "Factual Consistency Rate",
103
+ "metric_kind": "rate",
104
+ "metric_unit": "percent",
105
+ "metric_parameters": {},
106
+ "lower_is_better": false,
107
+ "score_type": "continuous",
108
+ "min_score": 0.0,
109
+ "max_score": 100.0,
110
+ "additional_details": {
111
+ "source_metric_key": "factual_consistency_rate",
112
+ "source_file": "microsoft/Phi-4/results_2025-12-10 14:57:16.944171.json",
113
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
114
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/microsoft/Phi-4/results_2025-12-10%2014%3A57%3A16.944171.json",
115
+ "evaluation_timestamp_source": "Parsed from the source result filename."
116
+ }
117
+ },
118
+ "score_details": {
119
+ "score": 96.3,
120
+ "details": {
121
+ "source_value_unit": "percent"
122
+ }
123
+ }
124
+ },
125
+ {
126
+ "evaluation_result_id": "vectara-hallucination-leaderboard/microsoft/Phi-4/1779880742.736885#answer_rate",
127
+ "evaluation_name": "Vectara Hallucination Leaderboard",
128
+ "source_data": {
129
+ "dataset_name": "vectara/hallucination-leaderboard",
130
+ "source_type": "url",
131
+ "url": [
132
+ "https://github.com/vectara/hallucination-leaderboard"
133
+ ]
134
+ },
135
+ "evaluation_timestamp": "2025-12-10 14:57:16.944171",
136
+ "metric_config": {
137
+ "evaluation_description": "Percentage of prompts for which the model produced an answer.",
138
+ "metric_id": "vectara-hallucination-leaderboard.answer_rate",
139
+ "metric_name": "Answer Rate",
140
+ "metric_kind": "rate",
141
+ "metric_unit": "percent",
142
+ "metric_parameters": {},
143
+ "lower_is_better": false,
144
+ "score_type": "continuous",
145
+ "min_score": 0.0,
146
+ "max_score": 100.0,
147
+ "additional_details": {
148
+ "source_metric_key": "answer_rate",
149
+ "source_file": "microsoft/Phi-4/results_2025-12-10 14:57:16.944171.json",
150
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
151
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/microsoft/Phi-4/results_2025-12-10%2014%3A57%3A16.944171.json",
152
+ "evaluation_timestamp_source": "Parsed from the source result filename."
153
+ }
154
+ },
155
+ "score_details": {
156
+ "score": 80.7,
157
+ "details": {
158
+ "source_value_unit": "percent"
159
+ }
160
+ }
161
+ },
162
+ {
163
+ "evaluation_result_id": "vectara-hallucination-leaderboard/microsoft/Phi-4/1779880742.736885#average_summary_length",
164
+ "evaluation_name": "Vectara Hallucination Leaderboard",
165
+ "source_data": {
166
+ "dataset_name": "vectara/hallucination-leaderboard",
167
+ "source_type": "url",
168
+ "url": [
169
+ "https://github.com/vectara/hallucination-leaderboard"
170
+ ]
171
+ },
172
+ "evaluation_timestamp": "2025-12-10 14:57:16.944171",
173
+ "metric_config": {
174
+ "evaluation_description": "Mean generated summary length in words; reported as a diagnostic metric.",
175
+ "metric_id": "vectara-hallucination-leaderboard.average_summary_length",
176
+ "metric_name": "Average Summary Length",
177
+ "metric_kind": "length",
178
+ "metric_unit": "words",
179
+ "metric_parameters": {},
180
+ "lower_is_better": false,
181
+ "score_type": "continuous",
182
+ "min_score": 0.0,
183
+ "max_score": 1000.0,
184
+ "additional_details": {
185
+ "source_metric_key": "average_summary_length",
186
+ "source_file": "microsoft/Phi-4/results_2025-12-10 14:57:16.944171.json",
187
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
188
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/microsoft/Phi-4/results_2025-12-10%2014%3A57%3A16.944171.json",
189
+ "evaluation_timestamp_source": "Parsed from the source result filename.",
190
+ "diagnostic_metric": "true",
191
+ "lower_is_better_note": "This is descriptive context rather than an optimization target."
192
+ }
193
+ },
194
+ "score_details": {
195
+ "score": 120.9,
196
+ "details": {
197
+ "source_value_unit": "words"
198
+ }
199
+ }
200
+ }
201
+ ]
202
+ }
data/vectara-hallucination-leaderboard/qwen/qwen3-8b/644ea67b-89da-4310-b05c-bb4064abc2ad.json ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vectara-hallucination-leaderboard/qwen/qwen3-8b/1779880742.736885",
4
+ "retrieved_timestamp": "1779880742.736885",
5
+ "evaluation_timestamp": "2025-12-10 14:57:15.832674",
6
+ "source_metadata": {
7
+ "source_name": "Vectara Hallucination Leaderboard",
8
+ "source_type": "documentation",
9
+ "source_organization_name": "Vectara",
10
+ "source_organization_url": "https://vectara.com",
11
+ "evaluator_relationship": "third_party",
12
+ "additional_details": {
13
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
14
+ "structured_results_hf_repo": "vectara/results",
15
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
16
+ "source_file": "qwen/qwen3-8b/results_2025-12-10 14:57:15.832674.json",
17
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/qwen/qwen3-8b/results_2025-12-10%2014%3A57%3A15.832674.json",
18
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
19
+ "underlying_evaluation_dataset": "Vectara Hallucination Leaderboard private evaluation dataset",
20
+ "underlying_evaluation_dataset_availability": "Private/proprietary dataset; not publicly released to avoid overfitting.",
21
+ "scoring_model": "Vectara HHEM-2.3",
22
+ "generation_temperature": "0 unless unavailable, per source documentation",
23
+ "evaluation_timestamp_source": "Parsed from the source result filename."
24
+ }
25
+ },
26
+ "model_info": {
27
+ "name": "qwen3-8b",
28
+ "id": "qwen/qwen3-8b",
29
+ "developer": "qwen",
30
+ "inference_platform": "dashscope",
31
+ "additional_details": {
32
+ "api_model_reference": "qwen3-8b",
33
+ "source_model_name": "qwen/qwen3-8b-",
34
+ "model_dtype": "float16",
35
+ "model_sha": "main",
36
+ "model_size": "small",
37
+ "accessibility": "open",
38
+ "deployment_type": "api"
39
+ }
40
+ },
41
+ "eval_library": {
42
+ "name": "unknown",
43
+ "version": "unknown",
44
+ "additional_details": {
45
+ "leaderboard_repository": "https://github.com/vectara/hallucination-leaderboard",
46
+ "structured_results_dataset": "https://huggingface.co/datasets/vectara/results",
47
+ "scoring_model": "Vectara HHEM-2.3"
48
+ }
49
+ },
50
+ "evaluation_results": [
51
+ {
52
+ "evaluation_result_id": "vectara-hallucination-leaderboard/qwen/qwen3-8b/1779880742.736885#hallucination_rate",
53
+ "evaluation_name": "Vectara Hallucination Leaderboard",
54
+ "source_data": {
55
+ "dataset_name": "vectara/hallucination-leaderboard",
56
+ "source_type": "url",
57
+ "url": [
58
+ "https://github.com/vectara/hallucination-leaderboard"
59
+ ]
60
+ },
61
+ "evaluation_timestamp": "2025-12-10 14:57:15.832674",
62
+ "metric_config": {
63
+ "evaluation_description": "Percentage of generated summaries judged to contain factual inconsistencies or unsupported claims.",
64
+ "metric_id": "vectara-hallucination-leaderboard.hallucination_rate",
65
+ "metric_name": "Hallucination Rate",
66
+ "metric_kind": "rate",
67
+ "metric_unit": "percent",
68
+ "metric_parameters": {},
69
+ "lower_is_better": true,
70
+ "score_type": "continuous",
71
+ "min_score": 0.0,
72
+ "max_score": 100.0,
73
+ "additional_details": {
74
+ "source_metric_key": "hallucination_rate",
75
+ "source_file": "qwen/qwen3-8b/results_2025-12-10 14:57:15.832674.json",
76
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
77
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/qwen/qwen3-8b/results_2025-12-10%2014%3A57%3A15.832674.json",
78
+ "evaluation_timestamp_source": "Parsed from the source result filename."
79
+ }
80
+ },
81
+ "score_details": {
82
+ "score": 4.8,
83
+ "details": {
84
+ "source_value_unit": "percent"
85
+ }
86
+ }
87
+ },
88
+ {
89
+ "evaluation_result_id": "vectara-hallucination-leaderboard/qwen/qwen3-8b/1779880742.736885#factual_consistency_rate",
90
+ "evaluation_name": "Vectara Hallucination Leaderboard",
91
+ "source_data": {
92
+ "dataset_name": "vectara/hallucination-leaderboard",
93
+ "source_type": "url",
94
+ "url": [
95
+ "https://github.com/vectara/hallucination-leaderboard"
96
+ ]
97
+ },
98
+ "evaluation_timestamp": "2025-12-10 14:57:15.832674",
99
+ "metric_config": {
100
+ "evaluation_description": "Percentage of generated summaries judged factually consistent.",
101
+ "metric_id": "vectara-hallucination-leaderboard.factual_consistency_rate",
102
+ "metric_name": "Factual Consistency Rate",
103
+ "metric_kind": "rate",
104
+ "metric_unit": "percent",
105
+ "metric_parameters": {},
106
+ "lower_is_better": false,
107
+ "score_type": "continuous",
108
+ "min_score": 0.0,
109
+ "max_score": 100.0,
110
+ "additional_details": {
111
+ "source_metric_key": "factual_consistency_rate",
112
+ "source_file": "qwen/qwen3-8b/results_2025-12-10 14:57:15.832674.json",
113
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
114
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/qwen/qwen3-8b/results_2025-12-10%2014%3A57%3A15.832674.json",
115
+ "evaluation_timestamp_source": "Parsed from the source result filename."
116
+ }
117
+ },
118
+ "score_details": {
119
+ "score": 95.2,
120
+ "details": {
121
+ "source_value_unit": "percent"
122
+ }
123
+ }
124
+ },
125
+ {
126
+ "evaluation_result_id": "vectara-hallucination-leaderboard/qwen/qwen3-8b/1779880742.736885#answer_rate",
127
+ "evaluation_name": "Vectara Hallucination Leaderboard",
128
+ "source_data": {
129
+ "dataset_name": "vectara/hallucination-leaderboard",
130
+ "source_type": "url",
131
+ "url": [
132
+ "https://github.com/vectara/hallucination-leaderboard"
133
+ ]
134
+ },
135
+ "evaluation_timestamp": "2025-12-10 14:57:15.832674",
136
+ "metric_config": {
137
+ "evaluation_description": "Percentage of prompts for which the model produced an answer.",
138
+ "metric_id": "vectara-hallucination-leaderboard.answer_rate",
139
+ "metric_name": "Answer Rate",
140
+ "metric_kind": "rate",
141
+ "metric_unit": "percent",
142
+ "metric_parameters": {},
143
+ "lower_is_better": false,
144
+ "score_type": "continuous",
145
+ "min_score": 0.0,
146
+ "max_score": 100.0,
147
+ "additional_details": {
148
+ "source_metric_key": "answer_rate",
149
+ "source_file": "qwen/qwen3-8b/results_2025-12-10 14:57:15.832674.json",
150
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
151
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/qwen/qwen3-8b/results_2025-12-10%2014%3A57%3A15.832674.json",
152
+ "evaluation_timestamp_source": "Parsed from the source result filename."
153
+ }
154
+ },
155
+ "score_details": {
156
+ "score": 99.9,
157
+ "details": {
158
+ "source_value_unit": "percent"
159
+ }
160
+ }
161
+ },
162
+ {
163
+ "evaluation_result_id": "vectara-hallucination-leaderboard/qwen/qwen3-8b/1779880742.736885#average_summary_length",
164
+ "evaluation_name": "Vectara Hallucination Leaderboard",
165
+ "source_data": {
166
+ "dataset_name": "vectara/hallucination-leaderboard",
167
+ "source_type": "url",
168
+ "url": [
169
+ "https://github.com/vectara/hallucination-leaderboard"
170
+ ]
171
+ },
172
+ "evaluation_timestamp": "2025-12-10 14:57:15.832674",
173
+ "metric_config": {
174
+ "evaluation_description": "Mean generated summary length in words; reported as a diagnostic metric.",
175
+ "metric_id": "vectara-hallucination-leaderboard.average_summary_length",
176
+ "metric_name": "Average Summary Length",
177
+ "metric_kind": "length",
178
+ "metric_unit": "words",
179
+ "metric_parameters": {},
180
+ "lower_is_better": false,
181
+ "score_type": "continuous",
182
+ "min_score": 0.0,
183
+ "max_score": 1000.0,
184
+ "additional_details": {
185
+ "source_metric_key": "average_summary_length",
186
+ "source_file": "qwen/qwen3-8b/results_2025-12-10 14:57:15.832674.json",
187
+ "source_commit": "7c104699e98ade53dd719f79ae9f7eb281c8107d",
188
+ "source_resolve_url": "https://huggingface.co/datasets/vectara/results/resolve/7c104699e98ade53dd719f79ae9f7eb281c8107d/qwen/qwen3-8b/results_2025-12-10%2014%3A57%3A15.832674.json",
189
+ "evaluation_timestamp_source": "Parsed from the source result filename.",
190
+ "diagnostic_metric": "true",
191
+ "lower_is_better_note": "This is descriptive context rather than an optimization target."
192
+ }
193
+ },
194
+ "score_details": {
195
+ "score": 83.6,
196
+ "details": {
197
+ "source_value_unit": "words"
198
+ }
199
+ }
200
+ }
201
+ ]
202
+ }