Question Answering on QA 8-objective
37.6EMPrompting
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PromptingAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 37.6 | 47.8 | 18.7 | 4.73 | 1.66 | |
| ACON UTAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 37.3 | 49.4 | 17.14 | 4.71 | 1.57 | |
| No compressionAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=None2025.10 | 36.6 | 48.8 | 15.78 | 10.35 | 3.32 | |
| ACON UTAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=Observation2025.10 | 36.4 | 47.5 | 16.33 | 4.97 | 1.28 | |
| LLMLinguaAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 36.3 | 48.1 | 17.68 | 5.68 | 2.24 | |
| ACON UTCOAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=Observation2025.10 | 33.6 | 46.1 | 14 | 4.22 | 0.81 | |
| ACON UTCOAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 33.5 | 45.8 | 17.79 | 4.65 | 1.5 | |
| RetrievalAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 33.1 | 43.8 | 20.06 | 5.11 | 2.62 | |
| LLMLinguaAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=Observation2025.10 | 32 | 41.4 | 14.23 | 5.16 | 1.35 | |
| FIFOAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=History2025.10 | 29.3 | 38.8 | 19.26 | 5.09 | 2.51 | |
| PromptingAgent=gpt-4.1, Compressor=gpt-4.1, Compression target=Observation2025.10 | 28.8 | 39.7 | 11.64 | 3.41 | 0.45 |