Explicit reasoning and standard RLVR training degrade LLM verification on subjective tasks, and a correctness-gated length reward recovers and exceeds baseline performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LLM Reasoning for Subjective Tasks: Failure Modes, Mitigation, and Dynamic Reasoning Routing
Explicit reasoning and standard RLVR training degrade LLM verification on subjective tasks, and a correctness-gated length reward recovers and exceeds baseline performance.