Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models
Chain-of-thought reasoning models such as DeepSeek-R1-Distill-Qwen-7B exhibit a bimodal convergence pattern: generations either terminate within a token budget (converged) or exhaust it without reaching a conclusion (non-converged). We characterize this phenomenon empirically, showing that converged generations achieve 90.3% accuracy on AIME 1983-2024 while non-converged ones achieve only 6.6%, with an overall convergence rate of 62.0%. We then ask whether this outcome is detectable early in the thinking chain using internal model representations. Training linear probes on hidden-state activat
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 53%Chain of Thought is a scaling trap. the next wave is latent reasoning (Coconut / HRM / RecrusiveMAS)... but then we hit the black box wall. Where does BDH fit? [D] →
- PossiblePossibly related (embedding) · 46%Exploring self-distilled reasoning for supervised fine-tuning with Amazon Nova →
- LinkedLinked via arxiv author · 85%Renuka Oladri →
“Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models”
- LinkedLinked via arxiv author · 85%Niveda Jawahar →
“Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models”
- LinkedLinked via arxiv author · 85%Abdirisak Mohamed →
“Token Budget Saturation and Mechanistic Early Detection of Reasoning Non-Convergence in Chain-of-Thought Models”