AI RESEARCH
Reasoning Shift: How Context Silently Shortens LLM Reasoning
arXiv CS.LG
•
ArXi:2604.01161v1 Announce Type: new Large language models (LLMs) exhibiting test-time scaling behavior, such as extended reasoning traces and self-verification, have nstrated remarkable performance on complex, long-term reasoning tasks. However, the robustness of these reasoning behaviors remains underexplored.