SOURCE-LINKED INTELLIGENCE

HSRM: Hidden-State Reward Models for Test-Time Verification

arXiv · AI, language, vision and robotics · article · Aug 31, 2026 · UTC

Large language models can often generate plausible mathematical reasoning traces, but reliably identifying the correct solution among multiple candidates remains a key challenge. Existing test-time reasoning pipelines typically rely on text-based verifiers that re-read each generated solution, making verification an expensive component of inference. Prior work has shown, however, that LLMs often encode correctness-related signals in their internal representations, including awareness of when their own answers are likely to be wrong. Building on this observation, we introduce HSRM, a lightweigh

Read original source ↗ Open in workspace

recordType

paper

region

Global