# When a language model is optimized for reasoning, does it still show embers of autoregression? An analysis of OpenAI o1

arXiv: https://arxiv.org/abs/2410.01792 · HTML: https://mlsanity.com/paper/2410.01792/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 4 of its 12 references that are in the map.
- First posted on arXiv 2 Oct 2024.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 4 of 4)

- [Sparks of Artificial General Intelligence: Early experiments with GPT-4](https://mlsanity.com/paper/2303.12712/) (2023)
- [Show Your Work: Scratchpads for Intermediate Computation with Language Models](https://mlsanity.com/paper/2112.00114/) (2021)
- [Embers of Autoregression: Understanding Large Language Models Through the Problem They are Trained to Solve](https://mlsanity.com/paper/2309.13638/) (2023)
- [Deciphering the Factors Influencing the Efficacy of Chain-of-Thought: Probability, Memorization, and Noisy Reasoning](https://mlsanity.com/paper/2407.01687/) (2024)

## Built on by (most cited first, 22 of 22)

- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [Towards Large Reasoning Models: A Survey of Reinforced Reasoning with Large Language Models](https://mlsanity.com/paper/2501.09686/) (2025)
- [Evaluating the Robustness of Analogical Reasoning in Large Language Models](https://mlsanity.com/paper/2411.14215/) (2024)
- [What is a Number, That a Large Language Model May Know It?](https://mlsanity.com/paper/2502.01540/) (2025)
- [Thinking beyond the anthropomorphic paradigm benefits LLM research](https://mlsanity.com/paper/2502.09192/) (2025)
- [A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments](https://mlsanity.com/paper/2509.25609/) (2025)
- [Evaluating GPT- and Reasoning-based Large Language Models on Physics Olympiad Problems: Surpassing Human Performance and Implications for Educational Assessment](https://mlsanity.com/paper/2505.09438/) (2025)
- [AlgoTrace: Algorithmic Primitives and Compositional Geometry of Reasoning in Language Models](https://mlsanity.com/paper/2510.15987/) (2025)
- [Beyond Statistical Learning: Exact Learning Is Essential for General Intelligence](https://mlsanity.com/paper/2506.23908/) (2025)
- [LogiDebrief: A Signal-Temporal Logic based Automated Debriefing Approach with Large Language Models Integration](https://mlsanity.com/paper/2505.03985/) (2025)
- [LLM Agents Are Hypersensitive to Nudges](https://mlsanity.com/paper/2505.11584/) (2025)
- [Evaluating Language Models' Evaluations of Games](https://mlsanity.com/paper/2510.10930/) (2025)
- [LINGOLY-TOO: Disentangling Reasoning from Knowledge with Templatised Orthographic Obfuscation](https://mlsanity.com/paper/2503.02972/) (2025)
- [ATTS: Asynchronous Test-Time Scaling via Conformal Prediction](https://mlsanity.com/paper/2509.15148/) (2025)
- [Large Language Models Develop Novel Social Biases Through Adaptive Exploration](https://mlsanity.com/paper/2511.06148/) (2025)
- [ContrastScore: Towards Higher Quality, Less Biased, More Efficient Evaluation Metrics with Contrastive Evaluation](https://mlsanity.com/paper/2504.02106/) (2025)
- [DWDP: Distributed Weight Data Parallelism for High-Performance LLM Inference on NVL72](https://mlsanity.com/paper/2604.01621/) (2026)
- [To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning](https://mlsanity.com/paper/2604.21632/) (2026)
- [Position: Behavioral Systems Require Behavioral Tests](https://mlsanity.com/paper/2608.18081/) (2026)
- [Towards Contamination Resistant Benchmarks](https://mlsanity.com/paper/2505.08389/) (2025)
- [Lifted Relational Probabilistic Inference via Implicit Learning](https://mlsanity.com/paper/2602.14890/) (2026)
- [Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play](https://mlsanity.com/paper/2606.19308/) (2026)

## Abstract (arXiv)

In "Embers of Autoregression" (McCoy et al., 2023), we showed that several large language models (LLMs) have some important limitations that are attributable to their origins in next-word prediction. Here we investigate whether these issues persist with o1, a new system from OpenAI that differs from previous LLMs in that it is optimized for reasoning. We find that o1 substantially outperforms previous LLMs in many cases, with particularly large improvements on rare variants of common tasks (e.g., forming acronyms from the second letter of each word in a list, rather than the first letter). Despite these quantitative improvements, however, o1 still displays the same qualitative trends that we observed in previous systems. Specifically, o1 -- like previous LLMs -- is sensitive to the probability of examples and tasks, performing better and requiring fewer "thinking tokens" in high-probability settings than in low-probability ones. These results show that optimizing a language model for reasoning can mitigate but might not fully overcome the language model's probability sensitivity.