2026· SemEval@ACL· pp. 3337-3342· 1 citation· 10 references
Computer Science
TL;DR
Through a rigorous three-stage empirical evolution and analysis of the diagnostic failure of an asymmetric logical ensemble, the experimental results demonstrate that epistemic noise suppression is strictly superior to heuristic recall compensation in abductive reasoning tasks.
Abstract
This paper describes a system for the SemEval-2026 Task 12 on Abductive Event Reasoning (AER). We systematically address the pervasive “over-selection” hallucination pathology in Instruction-tuned Large Language Models (LLMs), where generative models erroneously align distractor options with semantic priors rather than retrieved evidence. The proposed architecture utilizes a 32-billion-parameter foundational model within a Retrieval-Augmented Generation (RAG) pipeline. To combat the hallucination bottleneck and adapt to the strict penalty for incorrect predictions, we propose a Precision-Oriented Decoding (POD) strategy, which tightly couples low-temperature Zero-shot Chain-of-Thought (CoT) sampling with scaled marginalization (majority voting). De-ployed efficiently on consumer-grade hardware via Low-Rank Adaptation (LoRA) fine-tuning, our system achieved a highly competitive average score of 0.802 on the official test set. Through a rigorous three-stage empirical evolution and analysis of the diagnostic failure of an asymmetric logical ensemble, the experimental results demonstrate that epistemic noise suppression is strictly superior to heuristic recall compensation in abductive reasoning tasks.
Large vision-language models (LVLMs) frequently generate content unsupported by visual inputs. Preliminary experiments show that visual evidence is primarily incorporated into answer-side representations in early-to-middle decoder layers, while its direct influence progressively weakens in later layers. This attenuation suggests that visual evidence acquired earlier may be insufficiently utilized during subsequent generation. Based on this observation, we propose EviAnchor, a training-free and single-branch inference framework that preserves and reactivates visual evidence throughout generation. EviAnchor introduces Regional Evidence Anchor (REA) slots to progressively aggregate dense visual tokens into spatially structured representations. It then strengthens the current decision state's access to these visual anchors through decision-conditioned evidence routing, mitigating excessive dependence on textual context. Finally, the model resumes its native Transformer computation to integrate the retrieved visual evidence with question semantics and generation history. Experiments across POPE, CHAIR, and MMHal-Bench demonstrate consistent improvements in visual grounding.
Sihang Jia, Shuliang Liu, Song-Bo Yang et al.· 0 citations
The issue of hallucination detection for Retrieval-Augmented Generation (RAG) systems is an ongoing reliability challenge that arises with the scale up of LLM deployment. We introduce the HaloGuard: a systems-integration study that integrates three types of signals, multi-judge LLM consensus, uncertainty quantification through temperature sampling, and lexical text features, and evaluate with LibreEval (10,871 samples; GPT-4o, Claude-3.5-Sonnet, Llama-3.1-8B). They all rely on previous work; the novel aspect of the present paper is their principled combination, empirical decomposition through ablation, calibration analysis, and deployment characterization over operating thresholds, hallucination types, and six languages At a conservative threshold, HaloGuard achieves AUC-ROC = 0.986 [95% CI: 0.981-0.991] and F1 = 0.665 [0.639-0.690]; at the balanced max-F1 threshold, F1 = 0.769 (Precision = 0.812 / Recall = 0.731), directly addressing the precision-1.000 operating-point artifact. A fine-tuned variant (HaloGuard-FT, DeBERTa-v3-base) reaches F1 = 0.847 [0.831-0.863] and AUC = 0.994 [0.991-0.997]. Weight sensitivity analysis confirms F1 is stable across judge weight alpha in [0.5, 0.7] (F1 range <= 0.016), ruling out weight cherry-picking. Additional analyses quantify a 2.15x novelty differential (p < 0.001), source-attribution correlation (r = -0.319), and systematic cross-lingual F1 degradation (EN: 0.665 -> JA: 0.543). Calibration is well-behaved (ECE = 0.043). All results are scoped to LibreEval; cross-dataset validation on RAGTruth is the stated priority next step. Full judge prompts, hyperparameters, and code are documented for reproducibility.
Harshath R, G. S, S. Thangavel et al.· 2026 International Conferenc...· 0 citations
An integrated conceptual frame-work that couples attention- and perturbation-based explainability with lightweight hallucination-detection signals and token-efficient inference strategies is presented, and a set of cross-cutting consistency metrics are instrumented with a set of cross-cutting consistency metrics.
Sakshi Parate, Shreyans Sanyal· Advanced International Journ...· 0 citations
A concise two-axis framework that integrates an “intrinsic-extrinsic” distinction in source attribution introduced by Ji et al. with a “faithfulness-factuality” distinction in contextual grounding surveyed is presented, yielding four clearly defined hallucination types applicable across tasks, modalities and architectures.
Misbah Khan, Preston Billion-Polak, T. Khoshgoftaar· IEEE Access· 0 citations
The results recast package hallucination as both a measurement problem and a decoding-time control problem, and they demonstrate that the choice of defense must be matched to the threat model and recommendation utility.
Albérick Euraste Djiré, Iyiola E. Olatunji, Melissa Tessa et al.· 1 citation
The Latent Critic is introduced, a lightweight low-rank adapter that operates concurrently with a frozen base LLM's generation to actively restructure the transformer's residual stream---amplifying latent grounding signals and translating them into localized, natural language feedback within a single sequence.
S. Vijayvargiya, R. Lokesh· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
