Pith. sign in

REVIEW 5 cited by

Premise Order Matters in Reasoning with Large Language Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.08939 v3 pith:L6DVLFH6 submitted 2024-02-14 cs.AI cs.CL

classification cs.AIcs.CL
keywords reasoningorderingllmsorderpremiseperformanceaccuracybenchmark
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Large language models (LLMs) have accomplished remarkable reasoning performance in various domains. However, in the domain of reasoning tasks, we discover a frailty: LLMs are surprisingly brittle to the ordering of the premises, despite the fact that such ordering does not alter the underlying task. In particular, we observe that LLMs achieve the best performance when the premise order aligns with the context required in intermediate reasoning steps. For example, in deductive reasoning tasks, presenting the premises in the same order as the ground truth proof in the prompt (as opposed to random ordering) drastically increases the model's accuracy. We first examine the effect of premise ordering on deductive reasoning on a variety of LLMs, and our evaluation shows that permuting the premise order can cause a performance drop of over 30%. In addition, we release the benchmark R-GSM, based on GSM8K, to examine the ordering effect for mathematical problem-solving, and we again observe a significant drop in accuracy, relative to the original GSM8K benchmark.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Testing for LLM response differences: the case of a composite null consisting of semantically irrelevant query perturbations

    math.ST 2025-09 conditional novelty 6.0 of 10

    A new hypothesis test for binary LLM responses treats semantically equivalent query perturbations as an unknown null set and gives asymptotic validity and consistency guarantees under a uniformity assumption.

  2. Weak-to-Strong GraphRAG: Aligning Weak Retrievers with Large Language Models for Graph-based Retrieval Augmented Generation

    cs.CL 2025-06 conditional novelty 6.0 of 10

    ReG refines weak graph-retriever supervision with LLM-selected reasoning chains and reorganizes retrieved triples into coherent evidence chains, improving KGQA accuracy, data efficiency, and reasoning token efficiency.

  3. Fragile Preferences: A Deep Dive Into Order Effects in Large Language Models

    cs.AI 2025-06 unverdicted novelty 6.0 of 10

    LLMs show a quality-dependent position bias, favoring the first option for high-quality choices and later options for low-quality ones, and higher-temperature sampling can reveal the underlying preference.

  4. Generalised Probabilistic Modelling and Improved Uncertainty Estimation in Comparative LLM-as-a-judge

    cs.AI 2025-05 conditional novelty 6.0 of 10

    A generalised Product-of-Experts framework with a new 'probability of reordering' selection metric that reduces the number of LLM comparisons needed for ranking by about 50%.

  5. Towards High Supervised Learning Utility Training Data Generation: Data Pruning and Column Reordering

    cs.LG 2025-07 reject novelty 4.0 of 10

    PRRO combines signal-based data pruning and column reordering to improve the supervised learning utility of synthetic tabular data, but its evaluation is undermined by data manipulation and an ill-defined correlation measure.

Pith tools