LLM Reasoning for Subjective Tasks: Failure Modes, Mitigation, and Dynamic Reasoning Routing
This work introduces a conditional length-penalized post-training algorithm that intertwines verification accuracy with bounded reasoning length, halting collapse and recovering performance, and offers both a scalable algorithmic patch and a long-term architectural blueprint for aligning reasoning models with real-worl...
Jun-Cheng Dong, Ding Tong, Ishan Gupta et al.
· Proceedings of the 20th ACM... · 0 citations