mlsanityData to 14 Oct 2026

Uncovering Latent Chain of Thought Vectors in Language Models

arXiv 2409.14026 ↗First posted 21 Sep 2024cs.CL cs.AIcited by 14

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

In this work, we examine how targeted perturbations in the activation space of Language Models (LMs) can encode complex reasoning patterns. We inject steering vectors, derived from LM activations, into LMs during inference time and study whether these vectors can induce Chain-of-Thought (CoT) reasoning in LMs without the need for natural language prompting. We demonstrate this approach on Llama3 8B Instruct and Mistral 7B v0.2 Instruct and show that activation-space interventions achieve competitive, if not superior, performance compared to traditional CoT prompting across multiple reasoning benchmarks, including GSM8k, MMLU, AGI Eval, and ARC AI2. These findings suggest that neural network activations can encode reasoning patterns, offering a new application of activation space manipulation as a tool for tuning model behavior.

Key facts. Cited by 14 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 21 of its 27 references that are in the map. First posted on arXiv 21 Sep 2024.

Builds on

21 of 27 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Models28 Jan

Other references · 20

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  4. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  5. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  6. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  7. 2023A Survey of Large Language Modelscited by 2,576
  8. 2019Fine-Tuning Language Models from Human Preferencescited by 2,150
  9. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,986
  10. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  11. 2023Steering Llama 2 via Contrastive Activation Additioncited by 1,085
  12. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Modelscited by 886
8 more
  1. 2019Plug and Play Language Models: A Simple Approach to Controlled Text Generationcited by 799
  2. 2023Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedbackcited by 759
  3. 2023AGIEval: A Human-Centric Benchmark for Evaluating Foundation Modelscited by 716
  4. 2023The Internal State of an LLM Knows When It's Lyingcited by 659
  5. 2023In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steeringcited by 245
  6. 2023LongLoRA: Efficient Fine-tuning of Long-Context Large Language Modelscited by 244
  7. 2022Extracting Latent Steering Vectors from Pretrained Language Modelscited by 225
  8. 2024Beyond Accuracy: Evaluating the Reasoning Behavior of Large Language Models -- A Surveycited by 83

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 8; 2026: 6.

14 papers.

Sort:

By area:

  1. 2025Reasoning Beyond Language: A Comprehensive Survey on Latent Chain-of-Thought Reasoningcited by 78
  2. 2026The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlookcited by 43
  3. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Modelscited by 37
  4. 2026Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Modelscited by 27
  5. 2025Activation Steering for Chain-of-Thought Compressioncited by 24
  6. 2025Representation Engineering for Large-Language Models: Survey and Research Challengescited by 21
  7. 2026Do Latent-CoT Models Think Step-by-Step? A Mechanistic Study on Sequential Reasoning Taskscited by 12
  8. 2025Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinementcited by 10
  9. 2025Denoising Concept Vectors with Sparse Autoencoders for Improved Language Model Steeringcited by 9
  10. 2025The Impact of Off-Policy Training Data on Probe Generalisationcited by 9
  11. 2025Painless Activation Steering: An Automated, Lightweight Approach for Post-Training Large Language Modelscited by 2
  12. 2026Opening the Black Box: A Survey on the Mechanisms of Multi-Step Reasoning in Large Language Modelscited by 1
2 more
  1. 2026Revisiting Transformer Layer Parameterization Through Causal Energy Minimizationcited by 0
  2. 2026ALTSTEER: Selective Safety Steering for Moving Beyond Hard Refusals to Constructive Alternativescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Activation Control for Efficiently Eliciting Long Chain-of-thought Ability of Language Models
  2. 2026Triggering Chain-of-Thought via Latent Feature Interventions in Large Language Models
  3. 2025How does Chain of Thought Think? Mechanistic Interpretability of Chain-of-Thought Reasoning with Sparse Autoencoding
  4. 2024Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
  5. 2025Beyond Chains of Thought: Benchmarking Latent-Space Reasoning Abilities in Large Language Models
  6. 2025CoT Vectors: Transferring and Probing the Reasoning Mechanisms of LLMs
  7. 2024Chain-of-Thought Augmentation with Logit Contrast for Enhanced Reasoning in Language Models
  8. 2024On the Hardness of Faithful Chain-of-Thought Reasoning in Large Language Models