Circuit Claims Depend on What Is Extracted and How It Is Compared

• URL: http://arxiv.org/abs/2607.18921v1
• Date: Tue, 21 Jul 2026 10:04:29 GMT
• Title: Circuit Claims Depend on What Is Extracted and How It Is Compared
• Abstract summary: We argue that preserving behavior does not single out one circuit, because the claim it supports depends on which circuit is reported and how two circuits are compared.<n>We make this concrete in a synthetic Lean tactic-prediction benchmark.<n>A circuit-level claim is therefore well defined only once one states which circuit is reported, the pruning threshold used to extract it, and the level at which circuits are compared.
• Score: 10.373302428063091
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Circuit extraction identifies a small set of model components whose presence preserves a target behavior under ablation, and the resulting circuit is often read as the mechanism behind that behavior. We argue that this reading is under-determined: preserving behavior does not single out one circuit, because the claim it supports depends on which circuit is reported and how two circuits are compared. We make this concrete in a synthetic Lean tactic-prediction benchmark -- predicting the next step of a proof -- where fixed proof rules with randomized surface form let differences between extracted circuits be attributed to these choices rather than to the task. Across dense and weight-sparse checkpoints (most weights constrained to zero) of the same transformer, evaluated on atomic (single-rule) and compositional (multi-rule) proofs, we vary which extracted object is reported (a compact prediction-preserving circuit, a broader graph that also keeps surrounding read, write, and routing structure, or the smallest subgraph meeting a post-ablation loss threshold), and whether each attention head's query and key are represented jointly or separately. Exact component-to-component edge overlap is low and sensitive to these choices, at times dropping to a random baseline, while two coarser summaries stay stable: the set of selected attention heads, and the circuit-size ranking of conditions that differ in which supervised checkpoint initializes reinforcement learning (RL). The largest accuracy gains from RL on compositional proofs come with the most structure beyond the atomic circuits. A circuit-level claim is therefore well defined only once one states which circuit is reported, the pruning threshold used to extract it, and the level at which circuits are compared. We distill these requirements into a reporting practice for circuit-extraction studies.

Related papers

• Demystifying Variance in Circuit Discovery of LLMs [51.97489679448601]
  We develop a new circuit discovery method that substantially lessens resampling variance.<n>We show that rephrasing variance arises because prompts with different templates tend to activate different circuits in the model.<n>We argue that sparsity, which has been claimed to form more compact and interpretable task circuits, fails to solve this problem.
  arXiv  Detail & Related papers  (2026-06-15T16:25:07Z)
• Many Circuits, One Mechanism: Input Variation and Evaluation Granularity in Circuit Discovery [49.50794188352666]
  We show that structural differences between circuits are not sufficient evidence for distinct mechanisms.<n>Standard evaluation practice obscures this pattern, while edge-level evaluation reveals the many-to-one mapping from structure to function.
  arXiv  Detail & Related papers  (2026-06-04T15:10:14Z)
• Certified Circuits: Stability Guarantees for Mechanistic Circuits [80.30622018787835]
  Certified Circuits provides provable stability guarantees for circuit discovery.<n>On ImageNet and OOD datasets, certified circuits achieve up to 91% higher accuracy.
  arXiv  Detail & Related papers  (2026-02-26T13:07:31Z)
• Faster Probabilistic Error Cancellation [6.418044102466421]
  We propose a new method to perform PEC, which results in a lower sampling cost than the standard way.<n>We show the saving both analytically and numerically over the standard PEC.<n>We also demonstrated this method experimentally and found excellent agreement between the mitigated and the ideal values.
  arXiv  Detail & Related papers  (2025-06-04T21:38:56Z)
• Position-aware Automatic Circuit Discovery [59.64762573617173]
  We identify a gap in existing circuit discovery methods, treating model components as equally relevant across input positions.<n>We propose two improvements to incorporate positionality into circuits, even on tasks containing variable-length examples.<n>Our approach enables fully automated discovery of position-sensitive circuits, yielding better trade-offs between circuit size and faithfulness compared to prior work.
  arXiv  Detail & Related papers  (2025-02-07T00:18:20Z)
• CktGen: Specification-Conditioned Analog Circuit Generation [28.780603785886242]
  We introduce a task that directly generates analog circuits based on specified specifications. Specifically, we propose CktGen, a simple yet effective variational autoencoder (VAE) model. We conduct comprehensive experiments on the Open Circuit Benchmark (OCB) and introduce new evaluation metrics for cross-model consistency.
  arXiv  Detail & Related papers  (2024-10-01T18:35:44Z)
• Finding Transformer Circuits with Edge Pruning [71.12127707678961]
  We propose Edge Pruning as an effective and scalable solution to automated circuit discovery.<n>Our method finds circuits in GPT-2 that use less than half the number of edges compared to circuits found by previous methods.<n>Thanks to its efficiency, we scale Edge Pruning to CodeLlama-13B, a model over 100x the scale that prior methods operate on.
  arXiv  Detail & Related papers  (2024-06-24T16:40:54Z)
• Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms [35.514624827207136]
  Edge attribution patching (EAP), gradient-based approximation to interventions, has emerged as a scalable but imperfect solution to this problem. We introduce a new method - EAP with integrated gradients (EAP-IG) - that aims to better maintain a core property of circuits: faithfulness. Our experiments demonstrate that circuits found using EAP are less faithful than those found using EAP-IG, even though both have high node overlap with circuits found previously using causal interventions.
  arXiv  Detail & Related papers  (2024-03-26T15:44:58Z)
• Adaptive Planning Search Algorithm for Analog Circuit Verification [53.97809573610992]
  We propose a machine learning (ML) approach, which uses less simulations. We show that the proposed approach is able to provide OCCs closer to the specifications for all circuits.
  arXiv  Detail & Related papers  (2023-06-23T12:57:46Z)
• Qubit recycling and the path counting problem [0.0]
  Recently, it was shown that the qudits used in circuits of a convolutional form (e.g., Matrix Product State sand Multi-scaleanglement Renormalization Ansatz) can be reset unitarily. We analyze the fidelity of this protocol for a family of quantum circuits that interpolates between such circuits and local quantum circuits.
  arXiv  Detail & Related papers  (2023-01-09T23:59:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.