QA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
QA
42.5Mean PRR
109
QA
75.3ROC-AUC
64
QA Zero-shot Average
73.45QA Zero-shot Average
57
QA
3.66Speedup Factor
47
QA
1.73Average Rank
40
QA
49PRR
32
QA OOD StrQA SciQA
98.3StrQA Accuracy
28
QA (OBQA, ARC-E, ARC-C, CQA)
53.6OBQA Accuracy
20
QA
85.72Accuracy
18
QA ExAnte (test)
1.61d Leakage Rate
15
QA
63.31Performance Score
12
QA
70ASR Score (Before)
12
QA
59.5Accuracy
12
QA 8-objective
37.6EM
11
QA
62.5Steering Success
11
QA
117.17Throughput (tokens/s)
10
QA benchmarks
80.86ReCoRD Score
9
QA domain average
85.2Best Accuracy
8
QA
60.22HitRate@100
6
QA
75Win Rate
6
QA 12 languages
72.9Score
5
QA Qwen2-7B-Instruct v1 (test)
2.57Acceptance Length (τ)
4
QA XLow Tier
66R Score
3
QA Low Tier
76R (%)
3
QA High Tier
82.3R (%)
3