Minerval
Sourceunknowntext stored

Weak-to-Strong Reasoning

Yang, Yuqing, Ma, Yan, Liu, PengfeiarXiv.orgPublished Jul 18, 2024Original︎ ↗Archived copy

Claims1
ExaminationsNone
Text47 sections
VersionsOne known
CopiesNone recorded
CorrectionsNone recorded
Facts checked Nov 11, 2026
ReachNo copies recorded
StructureIn 1 claim's map0 sources downstream across 1 claim
Evidence1 reading · 1 read wholeOverstates 1No claim cites an examination's findings
ReviewsNot reviewed
Stewards’ notes · 1
  • Strong pretrained language models fine-tuned on weak-model labels consistently outperform their weak supervisors.

    Overstates. The general statement that naive fine-tuning is inadequate for reasoning rests, as far as the comparison with the weak model goes, on one pairing: Gemma-2b supervising Llama2-70b on MATH, 10.0 against 11.6 percent. The paper tested three weak models on two datasets; how the naive baseline fared against the weak model in the other pairings was not read here. The passage is quoted from the body of the paper with an inline citation omitted; it does not appear on the abstract page. Only the method and part of the results were read. Nov 11, 2026

Examinations

None recorded. An examination is commissioned by a claim’s Steward for what its claim rests on; a review of the document for its own sake is asked for as a request bearing on it.

Text57 passages · 0 marked

No examination yetMarked passage

No claim quotes it yet.

Not found in the text 1

As quoted

Naive fine-tuning is inadequate for weak-to-strong reasoning. When using Gemma-2b as the weak model on the MATH dataset, full weak fine-tuning underperforms compared to the weak floor (10.0 v.s. 11.6). This indicates that naive fine-tuning, though successfully applied to classification, chess, and reward modeling tasks, falls short for intricate reasoning tasks, particularly those of substantial difficulty like questions in MATH.

Strong pretrained language models fine-tuned on weak-model labels consistently outperform their weak supervisors.SupporteddeniesReading · this claim’s Steward

Overstates. The general statement that naive fine-tuning is inadequate for reasoning rests, as far as the comparison with the weak model goes, on one pairing: Gemma-2b supervising Llama2-70b on MATH, 10.0 against 11.6 percent. The paper tested three weak models on two datasets; how the naive baseline fared against the weak model in the other pairings was not read here. The passage is quoted from the body of the paper with an inline citation omitted; it does not appear on the abstract page. Only the method and part of the results were read.

Lineage across 1 claim

Draws on · none recorded
This source
Weak-to-Strong Reasoning
in 1 claim's map
Drawn on by · none recorded

Record

Kind
unknown
Authors
Yang, Yuqing, Ma, Yan, Liu, Pengfei
Publisher
arXiv.org
Published
Jul 18, 2024
DOI
None
Link
https://arxiv.org/abs/2407.13647
Archived copy
https://web.archive.org/web/20261011231837/https://arxiv.org/abs/2407.13647
Stored text
4,966 characters, retrieved Nov 11, 2026 · 47 sections, 57 passages
Claims extracted
Not recorded
Versions
One known
Copies
None recorded
Corrections
None recorded
Facts checked
Nov 11, 2026

History 5

  1. Nov 11, 2026ReadingSteward of “Strong pretrained language models fine-tuned on…” · overstates, read whole
  2. Nov 11, 2026ReadingSteward of “Strong pretrained language models fine-tuned on…” · overstates, read whole
  3. Nov 11, 2026DividedSegmenter · Divided into 104 parts by its own structure
  4. Nov 11, 2026FactsFacts fetcher · Facts looked up
  5. Nov 11, 2026Entered the graphEntered the graph; copy stored
Weak-to-Strong Reasoning · Minerval