初始化项目,由ModelHub XC社区提供模型
Model: Luimas/claim-extractor-detective-qwen3b Source: Original Platform
This commit is contained in:
16
benchmarks/base_benchmark.json
Normal file
16
benchmarks/base_benchmark.json
Normal file
@@ -0,0 +1,16 @@
|
||||
{
|
||||
"model": "unsloth/Qwen2.5-3B-Instruct-bnb-4bit (BASE, untrained)",
|
||||
"metrics": {
|
||||
"n": 15,
|
||||
"validity": 1.0,
|
||||
"avg_claims": 2.067,
|
||||
"avg_verification_questions": 3.71,
|
||||
"pct_claims_with_3plus_questions": 1,
|
||||
"contradiction_recall": 1.0,
|
||||
"date_extraction_rate": 1.0,
|
||||
"evidence_verbatim_rate": 1.0,
|
||||
"avg_claim_words": 5.548,
|
||||
"sarcasm_handling": 0.75,
|
||||
"avg_latency_s": 30.214
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user