Same Evidence, Different Target: Decoding How Diagnostic Evidence Bears on Causal Questions from Language-Model States

• URL: http://arxiv.org/abs/2607.26929v1
• Date: Wed, 29 Jul 2026 13:56:37 GMT
• Title: Same Evidence, Different Target: Decoding How Diagnostic Evidence Bears on Causal Questions from Language-Model States
• Abstract summary: We introduce paired prompts that repeat the same diagnostic evidence verbatim while changing the causal target.<n>Each prompt is labeled Favors, Challenges, Unresolved, or Wrong Target according to how the evidence bears on the causal question.<n>We analyze the final-token hidden state from the penultimate transformer block of Qwen2.5-7B-Instruct.
• Score: 7.032871370232397
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The same diagnostic result can support or challenge one causal claim yet fail to address another when the claims concern different populations, outcomes, estimands, pathways, or identifying assumptions. When the evidence and target vary together, a correct answer may reflect favorable or adverse wording, lexical overlap, or a familiar diagnostic pattern rather than matching the evidence to the causal question. We introduce paired prompts that repeat the same diagnostic evidence verbatim while changing the causal target. Each prompt is labeled Favors, Challenges, Unresolved, or Wrong Target according to how the evidence bears on the causal question. A pair is recovered only when both prompts are classified correctly. Using linear readouts trained on a separate development set, we analyze the final-token hidden state from the penultimate transformer block of Qwen2.5-7B-Instruct, Qwen3-8B, and Llama-3.1-8B-Instruct. On the 49-pair primary benchmark spanning nine diagnostic families, balanced accuracy ranges from 0.654 to 0.659 and 18-21 pairs are recovered. Two independent human reviewers assigned the same label to 95 of the 98 prompts (96.9%). Across checkpoints, balanced accuracy and complete-pair recovery exceed permutation nulls that preserve development scenario groups. In Qwen2.5, full-prompt balanced accuracy exceeds both restricted inputs, with paired-bootstrap intervals for both differences above zero. Readouts trained without development examples from the evaluated diagnostic family recover 21 pairs, including at least one in each of the nine families. The hidden-state readout exceeds a linear classifier on answer-option logits and text baselines in balanced accuracy and recovered pairs. These results show that the hidden state contains linearly decodable information about whether diagnostic evidence favors, challenges, or fails to address the causal target.

Related papers

• EviScope: Paired Counterfactual Evidence Diagnostics for Faithful and Efficient Grounded Language Models [0.0]
  Grounded language-model systems are often evaluated by final answer accuracy, yet a correct answer can be unsupported, drawn from the wrong source, or produced when evidence is insufficient or contradictory.<n>We introduce EviScope, a paired counterfactual benchmark that holds the question fixed while adding, removing, distracting, or contradicting its evidence.
  arXiv  Detail & Related papers  (2026-09-15T12:18:19Z)
• VeriDx: Earning the Right to Diagnose with Disease-Centric Verification [4.364186225903886]
  We introduce bfVeriDx, a disease-centric verification framework that links free-form diagnostic reasoning to structured disease profiles.<n>We instantiate VeriDx for complex respiratory diagnosis using guideline-derived disease profiles and expert-annotated longitudinal cases.
  arXiv  Detail & Related papers  (2026-09-12T16:05:07Z)
• DiagLoop: A Counterfactual Data Flywheel with Stage-Localized Reinforcement for Diagnostic LLMs [5.241146164724621]
  We present DiagLoop, a counterfactual data flywheel that converts codified physical relations or clinical guidelines into training supervision.<n>A training-only teacher proposes counterfactual worlds by varying causes, contexts, and observations, while an independent hybrid checker admits only valid worlds.<n>For nonterminal failures, a bounded repair probes downstream competence, and the resulting weakness profile guides subsequent data generation.
  arXiv  Detail & Related papers  (2026-08-04T13:49:41Z)
• Two Regimes of Chain-of-Thought Unfaithfulness: Behavioral Detection Fails Where Models Are Wrong [1.2744523252873352]
  Black-box (behavioral) detection of unfaithful CoT against FaithCoT-Bench's human annotations.<n>We find answer correctness structures the problem at every level.
  arXiv  Detail & Related papers  (2026-07-26T04:43:53Z)
• MedDDC-Eval: Diagnosis-Decoupled Evaluation of Multi-Turn Medical Consultation Agents [4.184241883998287]
  We introduce MedDDC-Eval, a diagnosis-decoupled evaluation testbed over held-out cases.<n>It applies the same frozen shared diagnostic reader to every policy-elicited history.<n>It reports diagnostic support, information-acquisition coverage, and efficiency.
  arXiv  Detail & Related papers  (2026-07-21T11:32:41Z)
• Before the Action: Benchmarking LLMs on Prospective Hypothesis Discovery [95.44432473367836]
  Large language models (LLMs) excel at answering pre-specified questions, yet their ability to navigate the open-ended, pre-conclusion stage of discovery remains largely unmeasured.<n>We introduce Prospective Hypothesis Discovery (PHD), which asks models to autonomously construct grounded, discnative, and testable hypothesis spaces from inconclusive evidence.<n>We introduce HypoArena, comprising HypoData, a benchmark of 988 cases across six scientific and analytical domains, and HypoEval, an evaluation framework for open-ended hypothesis sets.
  arXiv  Detail & Related papers  (2026-07-17T08:56:43Z)
• Code Correctness Signals in LLM Hidden States: Pre-Generation Probing and Repair Geometry [0.0]
  The correctness of the model's first-attempt code is linearly decodable from the prompt-final hidden state.<n>On 236 cleaned cases where the model attempts to repair a failed first attempt, the hidden state shift from the failing attempt to its repair carries a statistically detectable contrastive direction.
  arXiv  Detail & Related papers  (2026-06-12T15:06:15Z)
• CalBrief: A Pilot Diagnostic Benchmark for Evidence-Calibrated Scientific Briefing with Large Language Models [8.057966659934769]
  Large language models (LLMs) are increasingly used as research assistants, yet it remains unclear whether they can calibrate research takeaways to the strength and scope of the supporting evidence.<n>We study evidence-calibrated scientific briefing: given a bounded package of related papers, a system should generate package-level takeaways with evidence strength, scope boundaries, and missing-evidence caveats.<n>We use CalBrief, an auditable role/gap/strength framework, as a diagnostic probe to locate where briefing breaks down.
  arXiv  Detail & Related papers  (2026-06-11T06:49:52Z)
• MultiDx: A Multi-Source Knowledge Integration Framework towards Diagnostic Reasoning [66.94527468532843]
  We propose MultiDx, a two-stage diagnostic reasoning framework.<n>It first generates suspected diagnoses and reasoning paths by leveraging knowledge from web search, SOAP-formatted case, and clinical case database.<n>Then it integrates multi-perspective evidence through matching, voting, and differential diagnosis to generate the final prediction.
  arXiv  Detail & Related papers  (2026-04-27T08:46:29Z)
• Benchmarking Multi-turn Medical Diagnosis: Hold, Lure, and Self-Correction [72.89352076103889]
  Large language models (LLMs) achieve high accuracy in medical diagnosis when all clinical information is provided in a single turn.<n>We introduce MINT, a high-fidelity, multi-turn medical diagnosis benchmark comprising 1,035 cases.<n>We uncover three persistent behavioral patterns that significantly impact diagnostic decisions.
  arXiv  Detail & Related papers  (2026-04-06T00:23:10Z)
• MedEinst: Benchmarking the Einstellung Effect in Medical LLMs through Counterfactual Differential Diagnosis [13.241795322837861]
  We introduce MedEinst, a counterfactual benchmark with 5,383 paired clinical cases across 49 diseases.<n>We measure susceptibility via Bias Trap Rate--probability of misdiagnosing traps despite correctly diagnosing controls.
  arXiv  Detail & Related papers  (2026-01-10T17:39:25Z)
• Hide and Seek with LLMs: An Adversarial Game for Sneaky Error Generation and Self-Improving Diagnosis [51.88592148135258]
  We propose Hide and Seek Game (HSG), a dynamic adversarial framework for error generation and diagnosis.<n>HSG involves two adversarial roles: Sneaky, which "hides" by generating subtle, deceptive reasoning errors, and Diagnosis, which "seeks" to accurately detect them.<n> Experiments on several math reasoning tasks show that HSG significantly boosts error diagnosis, achieving 16.8%--31.4% higher accuracy than baselines like GPT-4o.
  arXiv  Detail & Related papers  (2025-08-05T12:45:21Z)
• Nonparametric Identifiability of Causal Representations from Unknown Interventions [63.1354734978244]
  We study causal representation learning, the task of inferring latent causal variables and their causal relations from mixtures of the variables. Our goal is to identify both the ground truth latents and their causal graph up to a set of ambiguities which we show to be irresolvable from interventional data.
  arXiv  Detail & Related papers  (2023-06-01T10:51:58Z)
• Towards Causality-Aware Inferring: A Sequential Discriminative Approach for Medical Diagnosis [142.90770786804507]
  Medical diagnosis assistant (MDA) aims to build an interactive diagnostic agent to sequentially inquire about symptoms for discriminating diseases. This work attempts to address these critical issues in MDA by taking advantage of the causal diagram. We propose a propensity-based patient simulator to effectively answer unrecorded inquiry by drawing knowledge from the other records.
  arXiv  Detail & Related papers  (2020-03-14T02:05:54Z)

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.