Attention Deficits in Language Models: Causal Explanations for Procedural Hallucinations

• URL: http://arxiv.org/abs/2602.19239v1
• Date: Sun, 22 Feb 2026 15:43:41 GMT
• Title: Attention Deficits in Language Models: Causal Explanations for Procedural Hallucinations
• Abstract summary: We study failure to execute a verifiable, prompt-grounded specification even when the correct value is present in context.<n>Specifically, failures decompose into Stage2A (gating) errors, where the model does not enter answer mode, and Stage2B (binding) errors, where it enters answer mode but selects the wrong candidate.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models can follow complex procedures yet fail at a seemingly trivial final step: reporting a value they themselves computed moments earlier. We study this phenomenon as \emph{procedural hallucination}: failure to execute a verifiable, prompt-grounded specification even when the correct value is present in context. In long-context binding tasks with a known single-token candidate set, we find that many errors are readout-stage routing failures. Specifically, failures decompose into Stage~2A (gating) errors, where the model does not enter answer mode, and Stage~2B (binding) errors, where it enters answer mode but selects the wrong candidate (often due to recency bias). In the hard regime, Stage~2B accounts for most errors across model families in our tasks (Table~1). On Stage~2B error trials, a linear probe on the final-layer residual stream recovers the correct value far above chance (e.g., 74\% vs.\ 2\% on Qwen2.5-3B; Table~2), indicating that the answer is encoded but not used. We formalize ``present but not used'' via available vs.\ used mutual information and pseudo-prior interventions, yielding output-computable diagnostics and information-budget certificates. Finally, an oracle checkpointing intervention that restates the true binding near the query can nearly eliminate Stage~2B failures at long distance (e.g., Qwen2.5-3B $0/400 \rightarrow 399/400$ at $k = 1024$; Table~8).

Related papers

• PluginEval: A Diagnostic Benchmark for Fine-Grained Error Attribution in Function Calling [11.1017602726296]
  Large Language Models increasingly operate as autonomous agents.<n>Current benchmarks face three structural limitations.<n>Eval is a benchmark constructed through a two-stage framework.
  arXiv  Detail & Related papers  (2026-08-09T13:25:54Z)
• Data and Evaluation Closed-Loop for Model Capability Enhancement [8.245706308161504]
  We show that evaluation can be routine, auditable, and experimentally validated rather than intuitive.<n>We show that the evaluation-to-data inference can be routine, auditable, and experimentally validated rather than intuitive.
  arXiv  Detail & Related papers  (2026-06-26T14:45:57Z)
• Cluster, Route, Escalate: Cascaded Framework for Cost-Aware LLM Serving [69.5155152169752]
  We propose a two-stage cascaded solution for large language models (LLMs)<n>Stage 1 clusters incoming queries and assigns each cluster to its most cost-effective model.<n>Stage 2 adds a quality estimation (QE) cascade; when an output from Stage 1 is judged low-quality, the query is escalated to a stronger model.<n>On the test datasets, the cascaded system retains 97-99% of the strongest model's accuracy while reducing Time Per Output Token (TPOT)<n>It requires only task-correctness labels and adapts to changes in the model pool without manual reconfiguration.
  arXiv  Detail & Related papers  (2026-06-25T18:29:24Z)
• Extreme Low-Bit Inference in Reasoning Models: Failure Modes and Targeted Recovery [40.94400211806987]
  We show that aggressive 2-bit inference can fail to deliver end-to-end speedup because instability in the generation process inflates total token count.<n>We analyze full reasoning traces of Qwen3 reasoning models across mathematical and commonsense benchmarks.<n>Our results show that extreme low-bit reasoning becomes practical when its failures are treated as controllable generation pathologies.
  arXiv  Detail & Related papers  (2026-06-01T10:04:09Z)
• Correction and Corruption: A Two-Rate View of Error Flow in LLM Protocols [51.56484100374058]
  We propose a paired-outcome measurement interface for auditing a single protocol step on exact-match tasks.<n>For each instance, the interface records a baseline correctness bit and a post-step correctness bit.<n>These rates predict accuracy changes and define a reusable empirical interface testable across seeds, mixtures, and pipelines.
  arXiv  Detail & Related papers  (2026-04-20T13:25:40Z)
• Did You Forget What I Asked? Prospective Memory Failures in Large Language Models [0.0]
  Large language models often fail to satisfy formatting instructions when they must simultaneously perform demanding tasks.<n>We study this behaviour through a prospective memory inspired lens from cognitive psychology, using a controlled paradigm.<n>Vulnerability is highly type-dependent, with drops up to 50%, while avoidance constraints remain comparatively robust.<n>A salience-enhanced format (explicit instruction framing plus a trailing reminder) recovers much of the lost compliance, restoring performance to 90-100% in many settings.
  arXiv  Detail & Related papers  (2026-03-07T05:58:19Z)
• Catch Your Breath: Adaptive Computation for Self-Paced Sequence Production [55.76222360698305]
  We explore a class of supervised training objectives that allow a language model to dynamically and autonomously scale the number of compute steps used for each input token.<n>For any token, the model can request additional compute steps by emitting a don't know> output.<n>We find that the CYB model requests additional steps when doing so improves accuracy, and the model adapts its processing time to token-level complexity and context.
  arXiv  Detail & Related papers  (2025-10-13T21:07:05Z)
• Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems [20.846301581161978]
  Failure attribution in multi-agent systems is a critical yet unsolved challenge.<n>Current methods treat this as a pattern recognition task over long conversation logs.<n>A2P Scaffolding transforms failure attribution from pattern recognition into a structured causal inference task.
  arXiv  Detail & Related papers  (2025-09-12T16:51:15Z)
• All is Not Lost: LLM Recovery without Checkpoints [0.1638581561083717]
  CheckFree is an efficient recovery method where a failing stage is substituted by a weighted average of the closest neighboring stages.<n>CheckFree and CheckFree+ outperform checkpointing and redundant computation in terms of convergence in wall-clock time by over 12%.
  arXiv  Detail & Related papers  (2025-06-18T13:48:33Z)
• Eliminating Hallucination-Induced Errors in LLM Code Generation with Functional Clustering [0.0]
  We present functional clustering, a black-box wrapper that eliminates nearly all hallucination-induced errors while providing a tunable confidence score.<n>Our verifier preserves baseline pass@1 on solvable tasks yet slashes the error rate of returned answers from 65% to 2%.<n>Because the method requires only sampling and sandbox execution, it applies unchanged to closed-source APIs and future models.
  arXiv  Detail & Related papers  (2025-05-16T18:19:38Z)
• Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling [90.86991492288487]
  evaluating constraint on every token can be prohibitively expensive.<n> LCD can distort the global distribution over strings, sampling tokens based only on local information.<n>We show that our approach is superior to state-of-the-art baselines.
  arXiv  Detail & Related papers  (2025-04-07T18:30:18Z)
• Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Models [38.79074982172423]
  We investigate the internal behavior of Transformer-based Large Language Models (LLMs) when they generate factually incorrect text. We propose modeling factual queries as constraint satisfaction problems. We find a strong positive relationship between the LLM's attention to constraint tokens and the factual accuracy of generations.
  arXiv  Detail & Related papers  (2023-09-26T17:48:55Z)
• Demonstrating a long-coherence dual-rail erasure qubit using tunable transmons [59.63080344946083]
  We show that a "dual-rail qubit" consisting of a pair of resonantly coupled transmons can form a highly coherent erasure qubit. We demonstrate mid-circuit detection of erasure errors while introducing $ 0.1%$ dephasing error per check. This work establishes transmon-based dual-rail qubits as an attractive building block for hardware-efficient quantum error correction.
  arXiv  Detail & Related papers  (2023-07-17T18:00:01Z)
• $k$NN Prompting: Beyond-Context Learning with Calibration-Free Nearest Neighbor Inference [75.08572535009276]
  In-Context Learning (ICL) formulates target tasks as prompt completion conditioned on in-context demonstrations. $k$NN Prompting first queries LLM with training data for distributed representations, then predicts test instances by simply referring to nearest neighbors. It significantly outperforms state-of-the-art calibration-based methods under comparable few-shot scenario.
  arXiv  Detail & Related papers  (2023-03-24T06:16:29Z)
• Towards Fine-Grained Information: Identifying the Type and Location of Translation Errors [80.22825549235556]
  Existing approaches can not synchronously consider error position and type. We build an FG-TED model to predict the textbf addition and textbfomission errors. Experiments show that our model can identify both error type and position concurrently, and gives state-of-the-art results.
  arXiv  Detail & Related papers  (2023-02-17T16:20:33Z)
• TACRED Revisited: A Thorough Evaluation of the TACRED Relation Extraction Task [80.38130122127882]
  TACRED is one of the largest, most widely used crowdsourced datasets in Relation Extraction (RE) In this paper, we investigate the questions: Have we reached a performance ceiling or is there still room for improvement? We find that label errors account for 8% absolute F1 test error, and that more than 50% of the examples need to be relabeled.
  arXiv  Detail & Related papers  (2020-04-30T15:07:37Z)

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.