Answer Generation on INSCIT
33.2F1 ScoreUniConv
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| UniConvLLM=Mistral-2-7b2026.01 | 33.2 | — | |
| OursLLM=Qwen-2.5-7b2026.01 | 30.1 | 58.5 | |
| ChatR1LLM=Qwen-2.5-3b2026.01 | 28.2 | — | |
| ChatR1LLM=Qwen-2.5-7b2026.01 | 27.8 | — | |
| EvoRAGLLM=Qwen-2.5-7b2026.01 | 26.8 | — | |
| ChatQALLM=LLaMA-3-8b2026.01 | 25.7 | — | |
| ClaudeLLM=Sonnet-3.52026.01 | 25 | — | |
| SFTLLM=Qwen-2.5-7b2026.01 | 24.5 | 53.9 | |
| SFTLLM=Qwen-2.5-3b2026.01 | 23.7 | 52.6 | |
| OursLLM=Qwen-2.5-3b2026.01 | 23.5 | 51.8 | |
| AgenticLMLLM=Qwen-3-8b2026.01 | 23.2 | 52.1 | |
| ChatGPTLLM=GPT-3.52026.01 | 22.6 | — | |
| AgenticLMLLM=Qwen-3-4b2026.01 | 21.6 | 50.8 | |
| ConvSearch-R1LLM=Qwen-2.5-3b2026.01 | 15.7 | 48.5 | |
| Search-R1LLM=Qwen-2.5-7b2026.01 | 9.1 | 57.7 | |
| Search-R1LLM=Qwen-2.5-3b2026.01 | 5.8 | 50.1 |