Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..

Premise Order Matters in Reasoning with Large Language Models

Authors: Xinyun Chen, Ryan Andrew Chi, Xuezhi Wang, Denny Zhou

ICML 2024 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We first examine the effect of premise ordering on deductive reasoning on a variety of LLMs, and our evaluation shows that...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.
Researcher Affiliation Collaboration 1Google Deep Mind 2Stanford University. Correspondence to: Xinyun Chen <EMAIL>, Ryan A. Chi <EMAIL>.
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper mentions releasing the R-GSM benchmark, which is a dataset, but does not state that the source code for their methodology is open-source or available.
Open Datasets Yes Besides logical reasoning, we construct R-GSM to further investigate the ordering effect on mathematical reasoning. Specifically, we build R-GSM on top of a subset of the GSM8K benchmark (Cobbe et al., 2021)...
Dataset Splits No The paper mentions using GSM8K test problems and generating logical reasoning problems, but does not provide explicit train/validation/test splits for reproducibility.
Hardware Specification No The paper does not specify any hardware used for the experiments (e.g., GPU/CPU models, memory).
Software Dependencies No The paper lists models used (e.g., GPT-4-turbo, PaLM 2-L) but does not provide specific version numbers for software dependencies or libraries.
Experiment Setup Yes We perform the greedy decoding with the temperature 0, and apply the zero-shot prompting in all experiments unless otherwise specified.