Download summaries/Ministral-8B-Instruct-2410_structured_extraction_report.txt from LatamBoard/leaderboard-results: direct link, hf CLI and curl.
- Browser
- Download file 4.37 kB
-
https://huggingface.co/datasets/LatamBoard/leaderboard-results/resolve/main/summaries/Ministral-8B-Instruct-2410_structured_extraction_report.txt
- Command line
-
hf download hf://datasets/LatamBoard/leaderboard-results/summaries/Ministral-8B-Instruct-2410_structured_extraction_report.txt
-
curl -L -o Ministral-8B-Instruct-2410_structured_extraction_report.txt https://huggingface.co/datasets/LatamBoard/leaderboard-results/resolve/main/summaries/Ministral-8B-Instruct-2410_structured_extraction_report.txt
4.37 kB
| ┌──────────────────────────────────────────────────────────────────────┐ | |
| │ Model: mistralai/Ministral-8B-Instruct-2410 │ | |
| │ Dataset: paraloq/json_data_extraction (476 samples) │ | |
| │ Date: 2025-10-18 17:56:34 │ | |
| └──────────────────────────────────────────────────────────────────────┘ | |
| OVERALL SCORE | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Extraction Quality Score (EQS): 0.842 ✅ GOOD | |
| 95% CI: [0.823, 0.862] | |
| PRIMARY METRICS | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| ✅ Schema Validity: 99.2% (472/476) Grade: A+ | |
| ⚠️ Field F1 (Partial): 0.764 Grade: B | |
| ❌ Exact Match: 32.1% (153/476) Grade: D | |
| ✅ Hallucination Rate: 8.1% Grade: A | |
| DETAILED ACCURACY | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Field F1 (Strict): 0.752 | |
| Field Precision: 0.774 | |
| Field Recall: 0.762 | |
| Type Accuracy: 88.6% | |
| MATCH DISTRIBUTION | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Exact 65.5% ██████████████████████████ | |
| Partial 2.6% █ | |
| Incorrect 12.3% ████ | |
| Missed 10.6% ████ | |
| Spurious 9.0% ███ | |
| COMPOSITE SCORE STATISTICS | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Mean: 0.893 | |
| Median: 1.000 | |
| StdDev: 0.273 | |
| Range: [0.001, 1.000] | |
| SAMPLE-LEVEL CONSISTENCY | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| EQS: μ=0.842, σ=0.217, range=[0.200, 1.000] | |
| F1: μ=0.764, σ=0.299, range=[0.000, 1.000] | |
| Consistency: Poor (highly variable) | |
| SCHEMA COMPLEXITY ANALYSIS | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Distribution by Complexity: | |
| Simple (442): EQS=0.838, F1=0.759, Halluc= 8.6% | |
| Medium ( 29): EQS=0.897, F1=0.839, Halluc= 2.9% | |
| Complex ( 1): EQS=1.000, F1=1.000, Halluc= 0.0% | |
| Correlations with Performance: | |
| complexity_score_vs_eqs r=+0.238 (weak positive) | |
| total_fields_vs_f1 r=+0.233 (weak positive) | |
| max_depth_vs_f1 r=+0.090 (negligible positive) | |
| PRODUCTION READINESS | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Quality Tier: Tier 2 (Good) | |
| Production Ready: ✅ YES | |
| Recommendation: DEPLOY WITH CONFIDENCE | |
| Model demonstrates strong extraction capability across | |
| the benchmark. Suitable for production use. | |
| PERFORMANCE | |
| ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ | |
| Throughput: 0.47 samples/second | |
| Total Duration: 1011.17 seconds | |
| Avg per Sample: 2.12 seconds | |