# Knowing Before Saying: LLM Representations Encode Information About Chain-of-Thought Success Before Completion

arXiv: https://arxiv.org/abs/2505.24362 · HTML: https://mlsanity.com/paper/2505.24362/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning.
- Builds on 15 of its 40 references that are in the map.
- First posted on arXiv 30 May 2025.

## Placed in

- LLM reasoning / [Prompting and chain-of-thought reasoning](https://mlsanity.com/lr/prompting-and-chain-of-thought-reasoning/)

## Builds on (most cited first, 15 of 15)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society](https://mlsanity.com/paper/2303.17760/) (2023)
- [Think before you speak: Training Language Models With Pause Tokens](https://mlsanity.com/paper/2310.02226/) (2023)
- [To CoT or not to CoT? Chain-of-thought helps mainly on math and symbolic reasoning](https://mlsanity.com/paper/2409.12183/) (2024)
- [Let's Think Dot by Dot: Hidden Computation in Transformer Language Models](https://mlsanity.com/paper/2404.15758/) (2024)
- [MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical Reasoning](https://mlsanity.com/paper/2310.03731/) (2023)
- [Text and Patterns: For Effective Chain of Thought, It Takes Two to Tango](https://mlsanity.com/paper/2209.07686/) (2022)
- [Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worse](https://mlsanity.com/paper/2410.21333/) (2024)
- [No Training Required: Exploring Random Encoders for Sentence Classification](https://mlsanity.com/paper/1901.10444/) (2019)
- [Language Modeling Teaches You More Syntax than Translation Does: Lessons Learned Through Auxiliary Task Analysis](https://mlsanity.com/paper/1809.10040/) (2018)
- [SVCCA: Singular Vector Canonical Correlation Analysis for Deep Learning Dynamics and Interpretability](https://mlsanity.com/paper/1706.05806/) (2017)
- [A Theoretical Understanding of Chain-of-Thought: Coherent Reasoning and Error-Aware Demonstration](https://mlsanity.com/paper/2410.16540/) (2024)
- [MyGO Multiplex CoT: A Method for Self-Reflection in Large Language Models via Double Chain of Thought Thinking](https://mlsanity.com/paper/2501.13117/) (2025)
- [Understanding Learning Dynamics Of Language Models with SVCCA](https://mlsanity.com/paper/1811.00225/) (2018)

## Built on by (most cited first, 23 of 23)

- [Can We Predict Alignment Before Models Finish Thinking? Towards Monitoring Misaligned Reasoning Models](https://mlsanity.com/paper/2507.12428/) (2025)
- [Beyond Token Probes: Hallucination Detection via Activation Tensors with ACT-ViT](https://mlsanity.com/paper/2510.00296/) (2025)
- [Silence the Judge: Reinforcement Learning with Self-Verifier via Latent Geometric Clustering](https://mlsanity.com/paper/2601.08427/) (2026)
- [The Value Axis: Language Models Encode Whether They're on the Right Track](https://mlsanity.com/paper/2606.17056/) (2026)
- [Early Stopping for Large Reasoning Models via Confidence Dynamics](https://mlsanity.com/paper/2604.04930/) (2026)
- [LCGuard: Latent Communication Guard for Safe KV Sharing in Multi-Agent Systems](https://mlsanity.com/paper/2605.22786/) (2026)
- [Doomed from the Start: Early Abort of LLM Agent Episodes via a Recall-Controlled Probe Cascade](https://mlsanity.com/paper/2607.06503/) (2026)
- [The Truth Lies Somewhere in the Middle (of the Generated Tokens)](https://mlsanity.com/paper/2605.09969/) (2026)
- [Attention as Binding: A Vector-Symbolic Perspective on Transformer Reasoning](https://mlsanity.com/paper/2512.14709/) (2025)
- [How Far Ahead Do LLMs Plan? Uncovering the Latent Horizon in Chain-of-Thought Reasoning](https://mlsanity.com/paper/2602.02103/) (2026)
- [Therefore I am. I Think](https://mlsanity.com/paper/2604.01202/) (2026)
- [Diagnosing Visual Ignorance in Vision-Language Models](https://mlsanity.com/paper/2606.06890/) (2026)
- [Reading Between the Dots: Decoding Hidden Computation across Filler Tokens](https://mlsanity.com/paper/2607.03502/) (2026)
- [Intra-request branch orchestration for efficient LLM reasoning](https://mlsanity.com/paper/2509.24957/) (2025)
- [Emergent Manifold Separability during Reasoning in Large Language Models](https://mlsanity.com/paper/2602.20338/) (2026)
- [Latent Programming Horizons in Coding Agents](https://mlsanity.com/paper/2607.05188/) (2026)
- [Architecture Determines Observability of Transformers](https://mlsanity.com/paper/2604.24801/) (2026)
- [Selective Safety Steering via Value-Filtered Decoding](https://mlsanity.com/paper/2605.14746/) (2026)
- [The Halt Vector: Internalizing a Causal Steering Intervention for Efficient Reasoning](https://mlsanity.com/paper/2608.28859/) (2026)
- [SKIP: a Self-knowledge-guided Step-wise Preference Learning Framework for Concise Reasoning](https://mlsanity.com/paper/2609.17019/) (2026)
- [The Commit-Abstain Circuit: Why Language Models Hallucinate Instead of Abstaining](https://mlsanity.com/paper/2609.32964/) (2026)
- [From Perception to Integration: Revisiting the Internal Dynamics of Reasoning in Vision-Language Models](https://mlsanity.com/paper/2609.34809/) (2026)
- [From Candidate Tokens to Final Outcomes: Decoding Relative Long-Horizon Candidate Value from Reasoning Model Representations](https://mlsanity.com/paper/2610.15412/) (2026)

## Abstract (arXiv)

We investigate whether the success of a zero-shot Chain-of-Thought (CoT) process can be predicted before completion. We discover that a probing classifier, based on LLM representations, performs well even before a single token is generated, suggesting that crucial information about the reasoning process is already present in the initial steps representations. In contrast, a strong BERT-based baseline, which relies solely on the generated tokens, performs worse, likely because it depends on shallow linguistic cues rather than deeper reasoning dynamics. Surprisingly, using later reasoning steps does not always improve classification. When additional context is unhelpful, earlier representations resemble later ones more, suggesting LLMs encode key information early. This implies reasoning can often stop early without loss. To test this, we conduct early stopping experiments, showing that truncating CoT reasoning still improves performance over not using CoT at all, though a gap remains compared to full reasoning. However, approaches like supervised learning or reinforcement learning designed to shorten CoT chains could leverage our classifier's guidance to identify when early stopping is effective. Our findings provide insights that may support such methods, helping to optimize CoT's efficiency while preserving its benefits.