LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations

• URL: http://arxiv.org/abs/2410.02707v3
• Date: Mon, 28 Oct 2024 12:33:44 GMT
• Title: LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations
• Authors: Hadas Orgad, Michael Toker, Zorik Gekhman, Roi Reichart, Idan Szpektor, Hadas Kotek, Yonatan Belinkov,
• Abstract summary: Large language models (LLMs) often produce errors, including factual inaccuracies, biases, and reasoning failures. Recent studies have demonstrated that LLMs' internal states encode information regarding the truthfulness of their outputs. We show that the internal representations of LLMs encode much more information about truthfulness than previously recognized.
• Score: 46.351064535592336
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) often produce errors, including factual inaccuracies, biases, and reasoning failures, collectively referred to as "hallucinations". Recent studies have demonstrated that LLMs' internal states encode information regarding the truthfulness of their outputs, and that this information can be utilized to detect errors. In this work, we show that the internal representations of LLMs encode much more information about truthfulness than previously recognized. We first discover that the truthfulness information is concentrated in specific tokens, and leveraging this property significantly enhances error detection performance. Yet, we show that such error detectors fail to generalize across datasets, implying that -- contrary to prior claims -- truthfulness encoding is not universal but rather multifaceted. Next, we show that internal representations can also be used for predicting the types of errors the model is likely to make, facilitating the development of tailored mitigation strategies. Lastly, we reveal a discrepancy between LLMs' internal encoding and external behavior: they may encode the correct answer, yet consistently generate an incorrect one. Taken together, these insights deepen our understanding of LLM errors from the model's internal perspective, which can guide future research on enhancing error analysis and mitigation.

Related papers

• MLLM can see? Dynamic Correction Decoding for Hallucination Mitigation [50.73561815838431]
  Multimodal Large Language Models (MLLMs) frequently exhibit hallucination phenomena. We propose a novel dynamic correction decoding method for MLLMs (DeCo) We evaluate DeCo on widely-used benchmarks, demonstrating that it can reduce hallucination rates by a large margin compared to baselines.
  arXiv  Detail & Related papers  (2024-10-15T16:57:44Z)
• Exploring Automatic Cryptographic API Misuse Detection in the Era of LLMs [60.32717556756674]
  This paper introduces a systematic evaluation framework to assess Large Language Models in detecting cryptographic misuses. Our in-depth analysis of 11,940 LLM-generated reports highlights that the inherent instabilities in LLMs can lead to over half of the reports being false positives. The optimized approach achieves a remarkable detection rate of nearly 90%, surpassing traditional methods and uncovering previously unknown misuses in established benchmarks.
  arXiv  Detail & Related papers  (2024-07-23T15:31:26Z)
• Analyzing LLM Behavior in Dialogue Summarization: Unveiling Circumstantial Hallucination Trends [38.86240794422485]
  We evaluate the faithfulness of large language models for dialogue summarization. Our evaluation reveals subtleties as to what constitutes a hallucination. We introduce two prompt-based approaches for fine-grained error detection that outperform existing metrics.
  arXiv  Detail & Related papers  (2024-06-05T17:49:47Z)
• LLMs' Reading Comprehension Is Affected by Parametric Knowledge and Struggles with Hypothetical Statements [59.71218039095155]
  Task of reading comprehension (RC) provides a primary means to assess language models' natural language understanding (NLU) capabilities. If the context aligns with the models' internal knowledge, it is hard to discern whether the models' answers stem from context comprehension or from internal information. To address this issue, we suggest to use RC on imaginary data, based on fictitious facts and entities.
  arXiv  Detail & Related papers  (2024-04-09T13:08:56Z)
• LLMs cannot find reasoning errors, but can correct them given the error location [0.9017736137562115]
  Poor self-correction performance stems from LLMs' inability to find logical mistakes, rather than their ability to correct a known mistake. We benchmark several state-of-the-art LLMs on their mistake-finding ability and demonstrate that they generally struggle with the task. We show that it is possible to obtain mistake location information without ground truth labels or in-domain training data.
  arXiv  Detail & Related papers  (2023-11-14T20:12:38Z)
• FactCHD: Benchmarking Fact-Conflicting Hallucination Detection [64.4610684475899]
  FactCHD is a benchmark designed for the detection of fact-conflicting hallucinations from LLMs. FactCHD features a diverse dataset that spans various factuality patterns, including vanilla, multi-hop, comparison, and set operation. We introduce Truth-Triangulator that synthesizes reflective considerations by tool-enhanced ChatGPT and LoRA-tuning based on Llama2.
  arXiv  Detail & Related papers  (2023-10-18T16:27:49Z)
• DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models [79.01926242857613]
  Large language models (LLMs) are prone to hallucinations, generating content that deviates from facts seen during pretraining. We propose a simple decoding strategy for reducing hallucinations with pretrained LLMs. We find that this Decoding by Contrasting Layers (DoLa) approach is able to better surface factual knowledge and reduce the generation of incorrect facts.
  arXiv  Detail & Related papers  (2023-09-07T17:45:31Z)
• Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts [21.34852490049787]
  We present the first comprehensive and controlled investigation into the behavior of large language models (LLMs) when encountering knowledge conflicts. We find that LLMs can be highly receptive to external evidence even when that conflicts with their parametric memory. On the other hand, LLMs also demonstrate a strong confirmation bias when the external evidence contains some information consistent with their parametric memory.
  arXiv  Detail & Related papers  (2023-05-22T17:57:41Z)

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.