mlsanityData to 14 Oct 2026

Therefore I am. I Think

arXiv 2604.01202 ↗First posted 1 Apr 2026cs.AIcited by 2

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

also in 1 more
Abstract (arXiv)

We consider the question: when a large language reasoning model makes a choice, did it think first and then decide to, or decide first and then think? In this paper, we present evidence that detectable, early-encoded decisions shape chain-of-thought in reasoning models. Specifically, we show that a simple linear probe successfully decodes tool-calling decisions from pre-generation activations with very high confidence, and in some cases, even before a single reasoning token is produced. Activation steering supports this causally: perturbing the decision direction leads to inflated deliberation, and flips behavior in many examples (between 7 - 79% depending on model and benchmark). We also show through behavioral analysis that, when steering changes the decision, the chain-of-thought process often rationalizes the flip rather than resisting it. Together, these results suggest that reasoning models can encode action choices before they begin to deliberate in text.

Key facts. Cited by 2 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 22 of its 33 references that are in the map. First posted on arXiv 1 Apr 2026.

Builds on

22 of 33 references.

Closest ancestors · 4

Same direction, newest first

  1. 2026Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought5 Mar
  2. 2025Knowing Before Saying: LLM Representations Encode Information About Chain-of-Thought Success Before Completion30 May
  3. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting7 May
  4. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Models28 Jan

Other references · 18

  1. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  2. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  3. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  4. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  5. 2023Steering Llama 2 via Contrastive Activation Additioncited by 1,085
  6. 2023Steering Language Models With Activation Engineeringcited by 922
  7. 2025Training Language Models to Reason Efficientlycited by 247
  8. 2024LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinationscited by 244
  9. 2024Programming Refusal with Conditional Activation Steeringcited by 206
  10. 2025Reasoning Models Know When They're Right: Probing Hidden States for Self-Verificationcited by 159
  11. 2023Future Lens: Anticipating Subsequent Tokens from a Single Hidden Statecited by 121
  12. 2025Thinkless: LLM Learns When to Thinkcited by 113
6 more
  1. 2023Improving Activation Steering in Language Models with Mean-Centringcited by 68
  2. 2024Monitoring Latent World States in Language Models with Propositional Probescited by 43
  3. 2025When2Call: When (not) to Call Toolscited by 42
  4. 2025The LLM Already Knows: Estimating LLM-Perceived Question Difficulty via Hidden Representationscited by 29
  5. 2025Measuring the Faithfulness of Thinking Drafts in Large Reasoning Modelscited by 23
  6. 2026Task-Specific Knowledge Distillation via Intermediate Probescited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Rotate2Think: Geometric Priming via Orthogonal Rotation to Improve Language Model Reasoningcited by 0
  2. 2026Committed Before Reasoning: Behavioral Reproduction and Preliminary Activation-Level Evidence of Answer Pre-Commitment in an Open-Weight LLMcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Reasoning aligns language models to human cognition
  2. 2025Rethinking the Chain-of-Thought: The Roles of In-Context Learning and Pre-trained Priors
  3. 2026When Steering Fails in Latent Reasoning: A Latent-to-Language Transition Gap
  4. 2026Legibility is Not Interpretability: Comparing Judged and Actual Importance in Chain-Of-Thought Reasoning
  5. 2026Post-Hoc Reasoning in Chain of Thought: Decoding and Steering Pre-Committed Answers
  6. 2026Emergent Search and Backtracking in Latent Reasoning Models
  7. 2024Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
  8. 2025Understanding and Steering the Cognitive Behaviors of Reasoning Models at Test-Time