Comprehensive Survey Maps LLM Reasoning Capabilities and Failure Modes Across 300+ Studies
Researchers published a systematic survey analyzing over 300 papers on how large language models perform reasoning tasks, organizing findings into a structured taxonomy covering nine reasoning paradigms. The survey identifies that while LLMs show progress in structured inference and multi-step problem solving, their reasoning remains inconsistent and sensitive to prompting strategies and model design. This work provides a unified framework for understanding current LLM reasoning limitations and guides future development of more robust reasoning systems.
A new arXiv preprint presents a comprehensive survey of LLM reasoning research, synthesizing insights from over 300 papers across multiple academic databases. The authors introduce a structured taxonomy organizing reasoning research into nine categories: Chain-of-Thought reasoning, multi-hop reasoning, mathematical reasoning, common sense reasoning, visual and temporal reasoning, code and algorithmic reasoning, retrieval-augmented reasoning, tool-augmented and agentic reasoning, and reinforcement learning-based reasoning. Beyond taxonomy, the survey analyzes methodological trends including prompting techniques, model architectures, training objectives, and evaluation benchmarks. The authors identify recurring failure modes such as reasoning hallucinations, brittle multi-step inference, weak causal abstraction, and poor cross-domain generalization. The work concludes by highlighting emerging research directions including meta-reasoning, self-evolving frameworks, multimodal reasoning, and socially grounded reasoning approaches.
What's missing
The survey's own limitations and scope boundaries are not detailed in the abstract—for instance, the temporal coverage of the 300+ papers reviewed, whether the taxonomy is exhaustive or representative, and any acknowledged gaps in the literature synthesis are not specified.
What different sources said
- arXiv cs.CLCenter
The Periodic Table of LLM Reasoning: A Structured Survey of Reasoning Paradigms, Methods, and Failure Modes
Related
Genetic Drift, Not Selection, Drives Rapid Feather Color Evolution in Island Bird Radiation
A new study of an island bird radiation found that rapid evolution of feather coloration is driven primarily by genetic drift in small populations rather than sexual or ecological selection. The research integrated whole-genome data with detailed plumage measurements across complete species sampling to test whether signaling trait evolution correlates with speciation rates. The findings suggest that neutral demographic processes play a central role in generating phenotypic diversity during island radiations, challenging assumptions about the mechanisms driving rapid evolution.
New AI Model Improves Prediction of Therapeutic Peptide Function from Protein Sequences
Researchers developed a lightweight CNN classifier that predicts whether peptide sequences have therapeutic properties, trained on a database of 54,655 peptides across 48 functional categories. The model uses a novel negative sampling strategy to reduce false positive rates from over 60% in previous approaches to 2.1%. This advancement could accelerate drug discovery by enabling faster computational screening of peptide candidates before expensive experimental testing.
Study Shows Different Metabolic Stress Models Produce Distinct Effects on Human Neuronal Networks
Researchers tested three common in vitro metabolic stress models on human-derived neuronal networks and found each produced different patterns of neuronal activity and cell damage. The models tested were hypoxia alone, oxygen-glucose deprivation (OGD), and hypoxia combined with glutamate exposure. The findings suggest that choice of experimental model significantly affects results and that combining electrophysiological and structural analyses is important for accurately assessing metabolic stress in stroke research.