Files
claim-extractor-detective-q…/benchmarks/teacher_benchmark.json

16 lines
416 B
JSON
Raw Normal View History

{
"model": "Qwen/Qwen2.5-14B-Instruct (TEACHER, no grammar)",
"metrics": {
"n": 15,
"validity": 0.933,
"avg_claims": 2.357,
"avg_verification_questions": 3.242,
"pct_claims_with_3plus_questions": 1,
"contradiction_recall": 1.0,
"date_extraction_rate": 1.0,
"evidence_verbatim_rate": 1.0,
"avg_claim_words": 7.333,
"sarcasm_handling": 1.0,
"avg_latency_s": 60.618
}
}