Study Examines How Post-Training Shapes Language Models' Verbalized Reasoning
A new arXiv paper investigates how post-training changes language models' use of internal information and their verbalized chain-of-thought reasoning, an area the authors say remains unclear.
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-11-11 05:00 · arXiv cs.CL
How Post-Training Influences Verbalized Algorithmic Computation in Language Models