Scholar Feed

LLM reasoning / chain-of-thought

Chain-of-Thought

Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

Heavily superseded#1 of 772 most-superseded · first seen Jan 28, 2022

Heavily superseded — a standard baseline that newer methods routinely beat

25 papers critique it · 13 beat it on benchmarks

What papers say

Verbatim critique sentences, each from a paper that cites Chain-of-Thought as a baseline.

errors can propagate through the reasoning chain, and there is no principled mechanism to revisit decisions or explore alternative strategies.
— STATe-of-Thoughts: Structured Action Templates for Tree-of-Thoughts
“However, despite these advances, LLMs frequently exhibit limitations in their logical consistency, accuracy, and self-correction abilities when confronted with highly intricate or chaotic reasoning problems”
— Enhancing Self-Correction in Large Language Models through Multi-Perspective Reflection
“However, CoT is fundamentally linear: once a reasoning step is generated, the model commits to it, often propagating early errors into final failures.”
— Tree of Thoughts as a Classical Heuristic Search Problem: Formal Foundations and Design Patterns
“Studies have shown that simple approaches like CoT are inadequate for tasks that demand decomposition into sub-tasks”
— GOT4Rec: Graph of Thoughts for Sequential Recommendation
“sprague2025to reported that on the Massive Multitask Language Understanding (MMLU) benchmark hendrycks2021measuring, 95\% of the performance gain from CoT is attributed to questions involving symbolic reasoning.”
— Does Self-Consistency Improve the Recall of Encyclopedic Knowledge?
“Nevertheless, CoT alone does not guarantee the factual correctness of the underlying statements within the reasoning chain.”
— Enhancing Factual Accuracy and Citation Generation in LLMs via Multi-Stage Self-Verification
“Chain-of-Thought (CoT) prompting~wei2022chain, zhang2023multimodal, lyu2023faithful, while enhancing reasoning, may not ensure its steps visually align with the image and can be sensitive to setup or resource-intensive.”
— MRFD: Multi-Region Fusion Decoding with Self-Consistency for Mitigating Hallucinations in LVLMs
“Although CoT has led to remarkable achievements, it does not always provide positive outcomes and sometimes hinders reasoning performance”
— Unveiling and Causalizing CoT: A Causal Pespective
“it still relies on a relatively simple, linear flow of thought, which can become limiting for tasks involving more complex reasoning”
— AURORA:Automated Training Framework of Universal Process Reward Models via Ensemble Prompting and Reverse Verification
“In this paper, we reveal a strikingly counterintuitive finding: Chain-of-Thought prompting unexpectedly degrades LLM performances in certain problem-solving contexts.”
— The Curse of CoT: On the Limitations of Chain-of-Thought in In-Context Learning
“this capability typically relies on explicit supervision (annotated CoT data), is confined to a discrete token space, and is ultimately capped by the base model's pre-trained capabilities”
— Pretraining with Token-Level Adaptive Latent Chain-of-Thought
“CoT's linear structure fails to capture the branching nature of mathematical reasoning, where multiple strategies are considered, partial arguments explored, and errors necessitate backtracking”
— Limits of PRM-Guided Tree Search for Mathematical Reasoning with LLMs

Beaten on benchmarks

Head-to-head results where a newer method reports beating Chain-of-Thought. Values are copied from the source paper's tables — verify against the cited paper.

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.