List your company

Question Answering on MSMARCO Out-of-distribution (test)

46.2AccuracyRLCR
33.61636.88340.1543.417May 15, 2026

Evaluation results

MethodLinks
RLCRBackbone=Qwen2.5-7B-Instruct2026.0546.28.5466.77
RLVRBackbone=Mistral-7B-Instruct2026.0543.145.6868.36
BaseBackbone=Mistral-7B-Instruct2026.0542.585.7368.52
CSRBackbone=Mistral-7B-Instruct2026.0542.544.8371.72
RLCRBackbone=Mistral-7B-Instruct2026.054215.9961.47
BaseBackbone=Llama-3.1-8B-Instruct2026.0541.538.7871.17
RLVRBackbone=Llama-3.1-8B-Instruct2026.0541.145.2671.38
RDBackbone=Llama-3.1-8B-Instruct2026.0540.916.5667.82
RLCRBackbone=Llama-3.1-8B-Instruct2026.0540.913.4763.13
CSRBackbone=Llama-3.1-8B-Instruct2026.0539.085.5571.72
BaseBackbone=Qwen2.5-7B-Instruct2026.0538.418.2268.04
RLVRBackbone=Qwen2.5-7B-Instruct2026.0538.398.3968.58
CSRBackbone=Qwen2.5-7B-Instruct2026.0536.745.571.34
RDBackbone=Qwen2.5-7B-Instruct2026.0536.715.3459.27
RDBackbone=Mistral-7B-Instruct2026.0534.124.3667.79