Question Answering on SciQ (Acc, nAUPC)
68AccuracyGround-Truth
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ground-TruthQA Model=T52026.04 | 68 | — | |
| T5 (beam)QA Model=T52026.04 | 64.7 | — | |
| TinyLlamaQA Model=T52026.04 | 63.8 | — | |
| SEALDomain=ID2026.04 | 62.6 | 40.2 | |
| T5 (contrast)QA Model=T52026.04 | 62.3 | — | |
| GPT-3 (clustering)QA Model=T52026.04 | 59.6 | — | |
| T5 (contrast)QA Model=BART2026.04 | 52.7 | — | |
| TinyLlamaQA Model=BART2026.04 | 52.6 | — | |
| Ground-TruthQA Model=BART2026.04 | 52.5 | — | |
| T5 (beam)QA Model=BART2026.04 | 51.7 | — | |
| GPT-3 (k-NN)QA Model=T52026.04 | 50.3 | — | |
| GPT-3 (COT-k-NN)QA Model=T52026.04 | 50.2 | — | |
| GPT-3 (clustering)QA Model=BART2026.04 | 47.7 | — | |
| GPT-3 (COT-k-NN)QA Model=BART2026.04 | 40.9 | — | |
| GPT-3 (k-NN)QA Model=BART2026.04 | 39.8 | — |