Long-context Question Answering on L-Eval
30.2Coursera QALongLLMLingua
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| LongLLMLinguaBackbone=Qwen3-8B, Input budget constraint=2,000-token, Tokens=2,196, 1/τ=4×2026.03 | 30.2 | 49 | 69.5 | 72.9 | 15.2 | 10.9 | 41.3 | |
| BEAVERBackbone=Qwen3-8B, Input budget constraint=2,000-token, Tokens=2,180, 1/τ=4×2026.03 | 28.3 | 57.4 | 73.4 | 80.3 | 17.7 | 13.4 | 45.1 | |
| LLMLingua-2Backbone=Qwen3-8B, Input budget constraint=2,000-token, Tokens=2,110, 1/τ=4×2026.03 | 25.3 | 53 | 74.2 | 78.4 | 16.9 | 12.8 | 43.5 | |
| LLMLinguaBackbone=Qwen3-8B, Input budget constraint=2,000-token, Tokens=2,122, 1/τ=4×2026.03 | 24.9 | 48 | 60.9 | 66.6 | 13.6 | 11.2 | 37.5 |