Explanation Multiplicity: Circuit-Level Interpretability Evidence Does Not Survive Defensible Analytic Variation

• URL: http://arxiv.org/abs/2608.13754v1
• Date: Thu, 13 Aug 2026 20:27:56 GMT
• Title: Explanation Multiplicity: Circuit-Level Interpretability Evidence Does Not Survive Defensible Analytic Variation
• Abstract summary: EU AI Act requires providers of high-risk systems to file technical documentation describing how the system reaches its decisions.<n>Mechanistic interpretability is the obvious source of such evidence, and circuit discovery is its most developed instrument.<n>We ask whether that evidence survives the condition under which it would be relied upon: two competent analysts, the same system, the same tool, different defensible settings.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The EU AI Act requires providers of high-risk systems to file technical documentation describing how the system reaches its decisions. Mechanistic interpretability is the obvious source of such evidence, and circuit discovery is its most developed instrument. We ask whether that evidence survives the condition under which it would be relied upon: two competent analysts, the same system, the same tool, different defensible settings. We pre-registered a crossed grid of seven analytic axes, every level taken from a published implementation, and mapped each discovered circuit through a deterministic claim map to a structured Annex IV statement. Across 15,840 pre-registered specifications on GPT-2 small and the indirect object identification task, of which 7,561 produced a claim, the derived statement flips across 73.2% of specification pairs (95% CI 0.725 to 0.738) and the modal claim commands 41.1% of the space. The evidence fails a filability criterion at every tolerance a conformity assessment body would plausibly accept. Standardising the single most influential choice, the evaluation metric, leaves the flip rate at 59.4%. Removing circuit size from the claim entirely and holding it fixed leaves 27.1% (95% CI 0.255 to 0.286), still above the pre-registered threshold. The circuits underlying these claims are structurally near-disjoint, median pairwise Jaccard overlap 4%, and functionally uncorrelated at Cohen's kappa 0.015, so the instability is not one mechanism described in different words. We give the filability criterion as a standalone protocol, and we report that one of the seven documented discovery objectives does not execute at all on the library's own canonical task. The study covers one model and one task, and whether the conclusion holds at scale is untested.

Related papers

• Evidence-Grounded Constraint Checking in Construction Documents [1.3787753513429983]
  We present an evidence-grounded pipeline that normalizes extracted facts and executes four-state rules deterministically.<n>We evaluate its PDF evidence allocator on 160 reference-based tasks from 29 construction projects.
  arXiv  Detail & Related papers  (2026-07-31T06:25:41Z)
• Looping Is Not Reliability: State-Bound Evidence and Typed Revision Contracts for Agentic Code Repair [36.56438114281786]
  We study the gap between finding a correct patch and retaining, verifying, and submitting it.<n>We derive an evidence-bound typed loop contract and instantiate its mechanically enforceable subset in a reference implementation.
  arXiv  Detail & Related papers  (2026-07-27T16:05:23Z)
• ITPEval: Benchmarking Formal Translation Across Interactive Theorem Provers [77.95933562142014]
  ITPEval is the first benchmark for evaluating automated formal proof translation across four major ITPs.<n>We release itpeval, a unified multi-ITP verification infrastructure with state-isolated warm backends.
  arXiv  Detail & Related papers  (2026-07-07T22:05:55Z)
• Beyond Refusal: A Same-Lineage Study of Aligned and Abliterated LLMs for Vulnerability Analysis [31.25049774595995]
  We study whether refusal behavior remains intact (Aligned) or has been refusalablated (Abliterated) within same-lineage models.<n>We evaluate Aligned and Abliterated states on vulnerability detection, CWE attribution, vulnerableline localization, root-cause localization, and executable patch validation.
  arXiv  Detail & Related papers  (2026-07-07T05:04:48Z)
• Output Type Before Quality: A Standards-Derived XAI Admissibility Rubric for Autonomous-Driving Safety [0.4958589793470846]
  Safety standards for ML-based autonomous driving specify the kind of evidence an assurance case must contain (directed cause-and-effect chains, quantified interventional effects, named root-cause variables)<n>Yet the XAI literature is organised by output type and technique family (saliency maps, feature attribution, counterfactuals, causal graphs, language traces)<n>We name this mismatch the evidence-type gap.<n>XAI method selection for safety assurance should be driven by lifecycle-stage evidence demand, not by method popularity.
  arXiv  Detail & Related papers  (2026-06-03T21:37:52Z)
• Let the Results Speak: A Replication-First Paradigm for LLM Behavioral Benchmarking [22.825786049667602]
  We propose a replication-first paradigm to anchor validity to a single human-rater consensus.<n>We certify the instrument via four properties -- reliability across K runs, cross-instrument replication across architecturally distinct judges, historical-footprint calibration via judges from earlier training cohorts, and pre-registered prediction.<n>We test it on emotional accompaniment by letting the self-evolve data-driven across: the dimensions are not pre-stipulated and the procedure stabilizes to a 9-dimension set.
  arXiv  Detail & Related papers  (2026-05-27T03:41:11Z)
• Distributional Energy-Based Models for Uncertainty-Aware Structured LLM Reasoning [40.342912574072024]
  Large Language Models produce structured outputs such as travel plans or code solutions.<n>Individual reasoning steps may appear correct while the output as a whole violates budgets, fails test cases, or contradicts earlier deductions.<n>We propose a learned quality scorer with deterministic analytical constraint penalties for verifying structured LLM outputs.
  arXiv  Detail & Related papers  (2026-05-15T17:08:27Z)
• PBT-Bench: Benchmarking AI Agents on Property-Based Testing [29.035258104995204]
  We introduce PBT-Bench, a benchmark of 100 curated property-based testing problems across 40 real Python libraries.<n>Each problem injects one or more semantic bugs (365 in total, mean 3.65 per problem) designed so that default-strategy random inputs almost never trigger them.<n>Bug recall under the PBT-guided prompt ranges from 42.1% to 83.4% across models; under the open-ended baseline, from 31.4% to 76.7%.
  arXiv  Detail & Related papers  (2026-05-13T18:01:05Z)
• The Alignment Tax: Response Homogenization in Aligned LLMs and Its Implications for Uncertainty Estimation [1.8345614451086532]
  RLHF-aligned language models exhibit response homogenization on TruthfulQA.<n>40-79% of questions produce a single semantic cluster across 10 i.i.d. samples.
  arXiv  Detail & Related papers  (2026-03-25T09:35:15Z)
• From Black Box to Glass Box: Cross-Model ASR Disagreement to Prioto Review in Ambient AI Scribe Documentation [43.148402136307716]
  Cross-model disagreement among heterogeneous ASR systems can act as a reference-free uncertainty signal.<n>We transcribed 50 publicly available medical audio clips with eight ASR systems spanning commercial APIs and open-source engines.<n>Low-agreement regions were enriched for content disagreements, with the content fraction increasing from 53.9% to 73.9% across quintiles of high-risk mass.
  arXiv  Detail & Related papers  (2026-03-02T13:02:13Z)

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.