Question Answering on HotpotQA (held-out)
81.2F1 ScoreRASER-3
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RASER-3Reader=Gemma-3-31B, N=2002026.06 | 81.2 | 1.8 | |
| RASER-2Reader=Gemma-3-31B, N=2002026.06 | 79.6 | 1.4 | |
| RASER-3Reader=GPT-OSS-120B, N=5002026.06 | 78.7 | 1.8 | |
| ONE-SHOT RAGReader=Gemma-3-31B, N=2002026.06 | 78.7 | 1.3 | |
| RASER-2Reader=GPT-OSS-120B, N=5002026.06 | 77.9 | 1.5 | |
| ONE-SHOT RAGReader=GPT-OSS-120B, N=5002026.06 | 77.7 | 1.4 | |
| RASER-3Reader=Mistral-S-119B, N=2002026.06 | 72.3 | 2.2 | |
| RASER-2Reader=Mistral-S-119B, N=2002026.06 | 70.2 | 1.4 | |
| ONE-SHOT RAGReader=Mistral-S-119B, N=2002026.06 | 69.9 | 1.3 | |
| RASER-3Reader=Llama-3-8B, N=2002026.06 | 65.5 | 2.2 | |
| RASER-3Reader=Llama-3.1-8B, N=2002026.06 | 65.4 | 2.4 | |
| RASER-3Reader=Phi-4-mini, N=2002026.06 | 65.2 | 2.2 | |
| RASER-2Reader=Llama-3-8B, N=2002026.06 | 63.5 | 1.3 | |
| RASER-2Reader=Phi-4-mini, N=2002026.06 | 63.5 | 1.4 | |
| ONE-SHOT RAGReader=Phi-4-mini, N=2002026.06 | 63.4 | 1.2 | |
| ONE-SHOT RAGReader=Llama-3-8B, N=2002026.06 | 63.2 | 1.2 | |
| RASER-2Reader=Llama-3.1-8B, N=2002026.06 | 62.4 | 1.5 | |
| ONE-SHOT RAGReader=Llama-3.1-8B, N=2002026.06 | 61.2 | 1.3 |