mlsanityData to 14 Oct 2026

Vernier: Probing Representational Misalignment Behind Lexical Gaps in Causal Reasoning

arXiv 2606.15733 ↗First posted 14 Jun 2026cs.CL cs.AIcited by 0

On the map: LLM reasoning → Causal reasoning in LLMs

Abstract (arXiv)

Instruction-tuned language models can answer the same causal-reasoning question differently after its English variable names are replaced by type-preserving placeholders, although the structural causal model and the gold answer are unchanged. We ask whether this lexical gap reflects information loss in the placeholder view or a misaligned read-out from a representation that still carries answer-relevant content. Vernier uses a paired-view weight update as an instrument and then inspects the mechanism left after the gap closes. In the working regimes, the evidence favours representational misalignment. A variable-name probe becomes more accurate on the placeholder view, and activation patching on Qwen-7B, Qwen-14B, and Llama-3.1-8B shows that the decision-token representation can transfer answer identity between views. The update that realigns the views is counterfactual augmentation over original and placeholder prompts, while the answer-subspace KL mainly sharpens intermediate answer-belief agreement. Success is bounded by model family, scale, and task. CRASS transfer is reliable across Qwen scales and Llama, e-CARE remains weak, and preliminary non-causal rename tasks show a similar qualitative pattern.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Causal reasoning in LLMs, one of 17 directions in LLM reasoning. Builds on 13 of its 43 references that are in the map. First posted on arXiv 14 Jun 2026.

Builds on

13 of 43 references.

Closest ancestors · 4

Same direction, newest first

  1. 2026Caliper: Probing Lexical Anchors versus Causal Structure in LLMs3 Jun
  2. 2026METER: Evaluating Multi-Level Contextual Causal Reasoning in Large Language Models13 Apr
  3. 2026CausalFlip: A Benchmark for LLM Causal Judgment Beyond Semantic Matching23 Feb
  4. 2025EconCausal: A Context-Aware Economic Reasoning Benchmark for Large Language Models8 Oct

Other references · 9

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2024Qwen2.5 Technical Reportcited by 4,942
  3. 2023Mistral 7Bcited by 3,164
  4. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,998
  5. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,917
  6. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  7. 2023Steering Language Models With Activation Engineeringcited by 922
  8. 2023Eliciting Latent Predictions from Transformers with the Tuned Lenscited by 623
  9. 2025An Investigation of Robustness of LLMs in Mathematical Reasoning: Benchmarking with Mathematically-Equivalent Transformation of Advanced Mathematical Problemscited by 8

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Convergence Without Understanding: When Language Models Agree on Representations but Disagree on Reasoning
  2. 2026Causal Tongue-Tie: LLMs Can Encode Causal Direction, But Their Yes/No Outputs Fail to Express
  3. 2026From Concept Alignment to Causal Grounding: An Intervention Test of Chain-of-Thought Faithfulness
  4. 2025Under the Shadow of Babel: How Language Shapes Reasoning in LLMs
  5. 2025What makes Reasoning Models Different? Follow the Reasoning Leader for Efficient Decoding
  6. 2025Answer-Centric or Reasoning-Driven? Uncovering the Latent Memory Anchor in LLMs
  7. 2024Nuance Matters: Probing Epistemic Consistency in Causal Reasoning
  8. 2024Calibrating Reasoning in Language Models with Internal Consistency