List your company

Instruction Following on ArenaHard

76.5Hard-Prompt Avg ScoreQUBRIC
23.77237.46151.1564.839Jun 2, 2026

Evaluation results

MethodLinks
QUBRIC2026.0676.558.9
RM-RLHFlength_penalty=1.02026.0671.442.3
SFT Baseline2026.067151.6
Rubric-RLquery=Original Query2026.0668.651.2
AutoIF-RLVR2026.0663.842.5
RM-RLHFlength_penalty=none2026.0625.821.4