inference-compute-reliably-improves-reasoning

OUT derived (depth 1)

Created 2026-06-21T13:13:49+00:00

Self-consistency (sampling multiple chain-of-thought paths and selecting by majority vote) and Tree of Thoughts (branching reasoning with backtracking) independently demonstrate that LLM reasoning reliability improves with inference-time compute investment — establishing a general compute-for-quality tradeoff at inference time — unless prompt sensitivity means each individual reasoning path is itself unreliable enough to undermine the ensemble.

Justifications

SL — Multiple strategies demonstrate inference-compute improves reasoning, contingent on individual paths being reliable enough for aggregation

Antecedents (all must be IN):

IN self-consistency-multiple-cot-majority-vote — Self-consistency improves Chain-of-Thought by sampling multiple reasoning paths and selecting the most common conclusion via majority vote (Wang et al., ICLR 2023, arXiv:2203.11171)
IN tree-of-thought-generalizes-cot — Tree-of-thought prompting generalizes chain-of-thought by exploring multiple reasoning paths in parallel with backtracking, using tree search algorithms (breadth-first, depth-first, beam search)

Unless (any of these IN defeats this justification):

IN prompt-sensitivity-40-percent-accuracy-shift — LLM performance is highly sensitive to prompt design, with accuracy shifts of over 40 percentage points from minor changes such as reordering examples, and up to 76 accuracy points difference across formatting changes