Self-contradictory Hallucinations of Large Language Models: Evaluation, Detection and Mitigation

• URL: http://arxiv.org/abs/2305.15852v3
• Date: Fri, 15 Mar 2024 21:04:34 GMT
• Title: Self-contradictory Hallucinations of Large Language Models: Evaluation, Detection and Mitigation
• Authors: Niels Mündler, Jingxuan He, Slobodan Jenko, Martin Vechev,
• Abstract summary: Large language models (large LMs) are susceptible to producing text that contains hallucinated content. We present a comprehensive investigation into self-contradiction for various instruction-tuned LMs. We propose a novel prompting-based framework designed to effectively detect and mitigate self-contradictions.
• Score: 5.043563227694139
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (large LMs) are susceptible to producing text that contains hallucinated content. An important instance of this problem is self-contradiction, where the LM generates two contradictory sentences within the same context. In this work, we present a comprehensive investigation into self-contradiction for various instruction-tuned LMs, covering evaluation, detection, and mitigation. Our primary evaluation task is open-domain text generation, but we also demonstrate the applicability of our approach to shorter question answering. Our analysis reveals the prevalence of self-contradictions, e.g., in 17.7% of all sentences produced by ChatGPT. We then propose a novel prompting-based framework designed to effectively detect and mitigate self-contradictions. Our detector achieves high accuracy, e.g., around 80% F1 score when prompting ChatGPT. The mitigation algorithm iteratively refines the generated text to remove contradictory information while preserving text fluency and informativeness. Importantly, our entire framework is applicable to black-box LMs and does not require retrieval of external knowledge. Rather, our method complements retrieval-based methods, as a large portion of self-contradictions (e.g., 35.2% for ChatGPT) cannot be verified using online text. Our approach is practically effective and has been released as a push-button tool to benefit the public at https://chatprotect.ai/.

Related papers

• ESPERANTO: Evaluating Synthesized Phrases to Enhance Robustness in AI Detection for Text Origination [1.8418334324753884]
  This paper introduces back-translation as a novel technique for evading detection. We present a model that combines these back-translated texts to produce a manipulated version of the original AI-generated text. We evaluate this technique on nine AI detectors, including six open-source and three proprietary systems.
  arXiv  Detail & Related papers  (2024-09-22T01:13:22Z)
• SparseCL: Sparse Contrastive Learning for Contradiction Retrieval [87.02936971689817]
  Contradiction retrieval refers to identifying and extracting documents that explicitly disagree with or refute the content of a query. Existing methods such as similarity search and crossencoder models exhibit significant limitations. We introduce SparseCL that leverages specially trained sentence embeddings designed to preserve subtle, contradictory nuances between sentences.
  arXiv  Detail & Related papers  (2024-06-15T21:57:03Z)
• DEMASQ: Unmasking the ChatGPT Wordsmith [63.8746084667206]
  We propose an effective ChatGPT detector named DEMASQ, which accurately identifies ChatGPT-generated content. Our method addresses two critical factors: (i) the distinct biases in text composition observed in human- and machine-generated content and (ii) the alterations made by humans to evade previous detection methods.
  arXiv  Detail & Related papers  (2023-11-08T21:13:05Z)
• Beyond Black Box AI-Generated Plagiarism Detection: From Sentence to Document Level [4.250876580245865]
  Existing AI-generated text classifiers have limited accuracy and often produce false positives. We propose a novel approach using natural language processing (NLP) techniques. We generate multiple paraphrased versions of a given question and inputting them into the large language model to generate answers. By using a contrastive loss function based on cosine similarity, we match generated sentences with those from the student's response.
  arXiv  Detail & Related papers  (2023-06-13T20:34:55Z)
• Towards a Robust Detection of Language Model Generated Text: Is ChatGPT that Easy to Detect? [0.0]
  This paper proposes a methodology for developing and evaluating ChatGPT detectors for French text. The proposed method involves translating an English dataset into French and training a classifier on the translated data. Results show that the detectors can effectively detect ChatGPT-generated text, with a degree of robustness against basic attack techniques in in-domain settings.
  arXiv  Detail & Related papers  (2023-06-09T13:03:53Z)
• DPIC: Decoupling Prompt and Intrinsic Characteristics for LLM Generated Text Detection [56.513637720967566]
  Large language models (LLMs) can generate texts that pose risks of misuse, such as plagiarism, planting fake reviews on e-commerce platforms, or creating inflammatory false tweets. Existing high-quality detection methods usually require access to the interior of the model to extract the intrinsic characteristics. We propose to extract deep intrinsic characteristics of the black-box model generated texts.
  arXiv  Detail & Related papers  (2023-05-21T17:26:16Z)
• ChatGPT as a Factual Inconsistency Evaluator for Text Summarization [17.166794984161964]
  We show that ChatGPT can evaluate factual inconsistency under a zero-shot setting. It generally outperforms previous evaluation metrics on binary entailment inference, summary ranking, and consistency rating. However, a closer inspection of ChatGPT's output reveals certain limitations including its preference for more lexically similar candidates, false reasoning, and inadequate understanding of instructions.
  arXiv  Detail & Related papers  (2023-03-27T22:30:39Z)
• SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models [55.60306377044225]
  "SelfCheckGPT" is a simple sampling-based approach to fact-check the responses of black-box models. We investigate this approach by using GPT-3 to generate passages about individuals from the WikiBio dataset.
  arXiv  Detail & Related papers  (2023-03-15T19:31:21Z)
• Verifying the Robustness of Automatic Credibility Assessment [50.55687778699995]
  We show that meaning-preserving changes in input text can mislead the models. We also introduce BODEGA: a benchmark for testing both victim models and attack methods on misinformation detection tasks. Our experimental results show that modern large language models are often more vulnerable to attacks than previous, smaller solutions.
  arXiv  Detail & Related papers  (2023-03-14T16:11:47Z)
• TextShield: Beyond Successfully Detecting Adversarial Sentences in Text Classification [6.781100829062443]
  Adversarial attack serves as a major challenge for neural network models in NLP, which precludes the model's deployment in safety-critical applications. Previous detection methods are incapable of giving correct predictions on adversarial sentences. We propose a saliency-based detector, which can effectively detect whether an input sentence is adversarial or not.
  arXiv  Detail & Related papers  (2023-02-03T22:58:07Z)
• TextHide: Tackling Data Privacy in Language Understanding Tasks [54.11691303032022]
  TextHide mitigates privacy risks without slowing down training or reducing accuracy. It requires all participants to add a simple encryption step to prevent an eavesdropping attacker from recovering private text data. We evaluate TextHide on the GLUE benchmark, and our experiments show that TextHide can effectively defend attacks on shared gradients or representations.
  arXiv  Detail & Related papers  (2020-10-12T22:22:15Z)

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.