claim
active
claim:chain-of-thought-makes-gsm8k-robust-to-ablation-because-the-model-externalizes-onto-the-page-what-it-would-otherwise-hold-in-the-j-spaceChain-of-thought makes GSM8K robust to ablation because the model externalizes onto the page what it would otherwise hold in the J-space.
Interpretation of ablation result.
Source paper
extracted_fromAuthors
Neighborhood — ranked by edge-count
Papers (1)
paper
Findings (1)
finding
- Externalization.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- GSM8K example: edited model produces identical chain of thought but miscalculates 260 as 220finding0.795Targeted arithmetic deficit.
- Analogy between LLM incoherence and schizophrenia symptoms
- Ablation result from Experiment 3 on chain-of-thought prompting effects.
- Finding from Study 2 showing reasoning models remain vulnerable under both prompting and activation steering
- Justifies SCR/TPP.
- under what conditions does chain-of-thought reflect genuine uncertainty resolution versus a learned performance?question0.751Key question addressed by the task difficulty analysis comparing MMLU and GPQA-Diamond
- Demonstrates robustness of inference stages to non-fixed-point limiting behavior