2026· IEEE Transactions on Information Forensics and Security· Vol 21, pp. 6605-6620· 0 citations· 45 references
Abstract
Large Language Models (LLMs) have demonstrated increasingly sophisticated capabilities in simulating human social interactions on social media platforms. Distinguishing LLMs from humans is essential for a deep analysis of online public opinion, while suffering from significant challenges, as existing approaches primarily rely on classifiers that learn different linguistic stylistic features of LLM-generated and human-written content. First, endless algorithmic confrontation games result in diminishing effectiveness against progressively more human-like LLM outputs. Second, for malignant events (such as fake news, rumor spread, etc.), relying on large-scale corpora to train classifiers is passive and lagging. To address these limitations, we propose PromptFishing, an active hallucination induction scheme that identifies LLMs through specific responses during interactions with social accounts. PromptFishing aims to embed machine-readable prompts into seemingly normal topics, creating a “semantic decoy” that triggers LLM hallucinations to produce a target response. To achieve this, we propose a gradient-guided two-stage beam search algorithm for adversarial suffix generation: the first stage seeks perturbations that deviate LLM responses from the original topic, while the second stage optimizes alignment with the target response. Experiments on data collected from Twitter show that our method achieves a true positive rate exceeding 0.85 while maintaining the false positive rate below 0.01. This work establishes a new paradigm for active LLM identification.
Large Language Models (LLMs) have achieved remarkable success in natural language generation but remain prone to hallucinations—generating content that is fluent but factually incorrect. While recent inference-time interventions like Contrastive Decoding (CD) effectively mitigate this by penalizing tokens favored by a "weak" hallucination-prone model, they introduce significant computational overhead (doubling inference latency) and fail to permanently align the model. In this paper, we propose \textbf{ContrastSFT}, a novel training framework to mitigate hallucinations in LLMs that internalizes the efficacy of contrastive decoding into the model's parameters via Contrastive Logit Regularization (CLR). Unlike standard Supervised Fine-Tuning (SFT) which indiscriminately maximizes the likelihood of ground-truth tokens, ContrastSFT dynamically recalibrates the training objective by subtracting the log-probabilities of a weak reference model. This effectively penalizes "easy" but potentially hallucinatory patterns captured by the weak model, forcing the model to learn more robust, factual representations. Extensive experiments on NLU benchmarks (ParaRel, WiCE) and Factuality tasks (HaluEval, MMLU) demonstrate that ContrastSFT achieves a 5-9\% absolute improvement over SFT and previous contrastive methods. Crucially, ContrastSFT eliminates the need for auxiliary models during deployment, retaining the high inference efficiency of standard LLMs. Code will be released.
Sifan Zhou· Poster Volume 0007 The 2026...· 0 citations
LLMs have become more widespread in many critical areas, which has increased the risk of hallucinations, or instances where generated outputs contain false or misleading information. To address this problem, this paper introduces HalluciNet v3.0+, which is an adaptive multi-LLM hallucination detection framework that employs advanced preprocessing, feature engineering, and classification methods for strong factual verification. In the preprocessing stage, the system begins with normalizing raw LLM-generated outputs; using linguistic parsing through spaCy to break up the LLM outputs into structured claims; and then using Transformer-based tokenization for contextual refinement of the structured claims. For the feature extraction step, the system uses both BERT and BGE embeddings to create dense semantic representations of the structured claims and the context and relationship dependencies between the structured claims and the retrieved evidence. A hybrid feature selection is achieved by combining Recursive Feature Elimination (RFE) and attention-weight optimization methods, which provides for the selection of the most discriminative features while minimizing dimensionality. A novel ensemble-based classification solution is implemented using DeBERTa-v3 for Natural Language Inference, XGBoost for structured decision learning, and consensus scoring across multiple LLM sources to produce a final predicted output. A two-step RAG (Retrieval-Augmented Generation) module optimizes retrieval and retrieval based on the quality of the retrieved item in a cross-encoder-based framework while preserving the inherent factual quality of the item's contents. This system is designed to be fully scalable and production-ready and is therefore expected to provide trustworthy enhancements to the use of LLM-driven products.
S.Rajeshwari, M. K, Mohamed Habith S et al.· International Conference Com...· 0 citations
Despite their widespread use, Large Language Models (LLMs) remain limited by a fundamental problem: the generation of plausible but false content, known as hallucinations. Most existing detection methods operate at the answer or sentence level, yet per-token detection is essential for localizing hallucinated spans and enabling fine-grained interventions. In this paper, we explore the use of the Mixture-of-Experts (MoE) paradigm to address this gap. In MoE architectures, a single forward pass activates a sparse subset of experts (i.e., distinct feedforward networks per layer) via a routing mechanism, producing internal signals (e.g., router entropy, expert disagreement, and expert usage patterns) that are unavailable in dense architectures and have not been previously exploited for hallucination detection. To this end, we introduce InnerExpert, the first method to leverage these MoE-specific signals for per-token hallucination detection. InnerExpert combines routing-level and standard transformer signals into compact per-token feature vectors, classified by a lightweight detector trained on labels produced by an LLM-as-a-judge pipeline, which enables continuous model updates without manual annotation. Our results show that InnerExpert outperforms existing methods across five datasets and two MoE architectures, achieving up to 0.91 answer-level and 0.76 token-level AUROC, while requiring only a single forward pass.
João Fonseca, Rodrigo Rodrigues, Paolo Romano· 0 citations
This work proposes CounterfactualLVLM, a training-free and plug-and-play framework that mitigates object hallucinations via small-model-assisted counterfactual reasoning and highlights the power of counterfactual guidance as a simple yet effective paradigm for enhancing factual grounding in LVLM-based multi-modal reasoning.
Xilin Li, Boyue Wang, Xiaoqian Ju et al.· Multimedia Systems· 0 citations
This work studies hallucination from the perspective of dynamic representation shift during generation and proposes an online projection-based intervention on intermediate hidden states to suppress the hallucination-related directions, mitigating hallucinations while preserving language quality.
SeeMe is proposed, a training-free framework that introduces the concept of feature engineering from traditional machine learning into LVLMs and restructures visual tokens through a three-stage token engineering process to suppress hallucination sources while preserving informative visual evidence.
Kai Tang, Jinhao You, Bohua Zhang et al.· 2 citations