List your company

Long-form Factual Generation on LongFact

84.2Fact Recall (FR) - ScienceArbGraph
58.61665.25871.978.542Apr 20, 2026

Evaluation results

MethodLinks
ArbGraphBackbone=LLaMA-3.1-8B2026.0484.285.186.185.3848385.984.784.9
MCTS-RAGBackbone=LLaMA-3.1-8B2026.0483.881.68482.580.581.977.181.981.7
ArbGraphBackbone=Qwen3-4B2026.0483.584.486.283.382.579.884.781.683.3
MCTS-RAGBackbone=Qwen3-4B2026.0480.477.485.275.882.479.276.377.979.2
FIREBackbone=Qwen3-4B2026.0479.776.682.877.180.279.17481.278.8
FIREBackbone=LLaMA-3.1-8B2026.0479.278.480.577.176.275.178.273.678.3
CRITICBackbone=Qwen3-4B2026.0478.479.782.473.376.675.269.375.476.3
CRITICBackbone=LLaMA-3.1-8B2026.0478.269.278.171.370.174.369.275.273.3
ReActBackbone=LLaMA-3.1-8B2026.0469.569.271.968.767.368.366.870.469
CoTBackbone=LLaMA-3.1-8B2026.0469.156.168.361.966.66973.460.864.7
CoTBackbone=Qwen3-4B2026.0465.77065.662.555.258.555.869.463.1
ReActBackbone=Qwen3-4B2026.0465.659.271.362.464.46650.670.863.3
Standard RAGBackbone=Qwen3-4B2026.0464.973.764.761.658.556.756.668.663.5
Standard RAGBackbone=LLaMA-3.1-8B2026.0459.668.869.86662.259.557.772.164.6