REVIEW 5 cited by
Deciphering the Factors Influencing the Efficacy of Chain-of-Thought: Probability, Memorization, and Noisy Reasoning
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
read the original abstract
Chain-of-Thought (CoT) prompting has been shown to enhance the multi-step reasoning capabilities of Large Language Models (LLMs). However, debates persist about whether LLMs exhibit abstract generalization or rely on shallow heuristics when given CoT prompts. To understand the factors influencing CoT reasoning we provide a detailed case study of the symbolic reasoning task of decoding shift ciphers, where letters are shifted forward some number of steps in the alphabet. We analyze the pattern of results produced by three LLMs -- GPT-4, Claude 3, and Llama 3.1 -- performing this task using CoT prompting. By focusing on a single relatively simple task, we are able to identify three factors that systematically affect CoT performance: the probability of the task's expected output (probability), what the model has implicitly learned during pre-training (memorization), and the number of intermediate operations involved in reasoning (noisy reasoning). We show that these factors can drastically influence task accuracy across all three LLMs; e.g., when tested with GPT-4, varying the output's probability of occurrence shifts accuracy from 26% to 70%. Overall, we conclude that CoT prompting performance reflects both memorization and a probabilistic version of genuine reasoning. Code and data at this https://github.com/aksh555/deciphering_cot
Forward citations
Cited by 5 Pith papers
-
Task Complexity Matters: An Empirical Study of Reasoning in LLMs for Sentiment Analysis
Reasoning in LLMs helps only for complex 27-class emotion recognition and hurts simple binary sentiment, across seven model families and 504 configurations.
-
Evaluating the Robustness of Analogical Reasoning in Large Language Models
GPT models solve original analogy tasks but fail many simple variants that humans handle easily, showing their analogy performance is not robust.
-
Ask-Before-Detection: Identifying and Mitigating Conformity Bias in LLM-Powered Error Detector for Math Word Problem Solutions
LLM error detectors favor conventional solution formats; generating an adaptive reference solution before grading mitigates this conformity bias on a 200-example GSM8K subset.
-
Understanding and Benchmarking Artificial Intelligence: OpenAI's o3 Is Not AGI
A conceptual analysis arguing that OpenAI's o3 solves ARC-AGI by brute-force search over predefined operations, so its high score is not evidence of AGI, and proposes a new definition and benchmark for intelligence.
-
Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
Chain-of-thought prompting makes LLMs imitate exemplar formatting, concentrate their final answer probabilities, and activate a wider set of final-layer neurons.
Discussion (0). Continue with ORCID to comment.