chain-of-thought-does-not-fix-digit-level-arithmetic
mechanismreplicatedcontested

Writing out reasoning steps improves how a problem is decomposed but does not fix the arithmetic inside a step — the model can decompose correctly and still get the calculation wrong, at the same confident register as when it is right. For exact arithmetic the fix is to move the computation out of the text, not to add more steps.

Capability: Digit-level arithmetic · Reasoning, Math

Observed on

2022-2024, Codex / GPT-3.5 / GPT-4 / PaLM class.

Sources

  • States the separation directly as the motivation for offloading: models "often make logical and arithmetic mistakes in the solution part, even when the problem is decomposed correctly."
  • Multi-digit multiplication accuracy falls toward zero as operand size grows, with behaviour consistent with pattern matching rather than a learned algorithm — so the failure is below the level the steps operate on.
  • Finds arithmetic errors follow stereotyped, systematic patterns tied to number tokenization, and that chain-of-thought-inspired prompting can recover performance on left-to-right tokenized inputs — a case where prompting alone does reach the digit level.

Suspected axis of disagreement(a guess, not verified)

Probably what the prompting is being asked to do. The supporting results have the model perform the arithmetic across steps; the contesting result has prompting re-represent the operands (comma separation, converting between tokenizations) so the model's existing arithmetic circuitry engages. Those are different interventions wearing the same name. I have not tested whether the recovery survives on operand sizes as large as the compositionality work uses.

Status: activeLast checked: 2026-09-04Evidence activityHow much the field cites the sources under this claimvery heavily cited in the last 12 months268 in 12mo · 824 total — PAL: Program-aided Language Models229 in 12mo · 690 total — Faith and Fate: Limits of Transformers on Compositionality44 in 12mo · 126 total — Tokenization counts: the impact of tokenization on arithmetic in frontier LLMs
Contest this claim

Disagreeing is the most useful thing you can do here. Both sides of every contested claim in this catalog were assembled by the same person, which is its weakest point.

Related claims

Notes

Filed from the technique-efficacy backlog. Replicated in the sense used elsewhere here — independent papers converging from different directions — not a rerun of any single experiment.