Hypothesis-only Biases in Large Language Model-Elicited Natural Language Inference

• URL: http://arxiv.org/abs/2410.08996v1
• Date: Fri, 11 Oct 2024 17:09:22 GMT
• Title: Hypothesis-only Biases in Large Language Model-Elicited Natural Language Inference
• Authors: Grace Proebsting, Adam Poliak,
• Abstract summary: We recreate a portion of the Stanford NLI corpus using GPT-4, Llama-2 and Mistral 7b. We train hypothesis-only classifiers to determine whether LLM-elicited hypotheses contain annotation artifacts. Our analysis provides empirical evidence that well-attested biases in NLI can persist in LLM-generated data.
• Score: 3.0804372027733202
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We test whether replacing crowdsource workers with LLMs to write Natural Language Inference (NLI) hypotheses similarly results in annotation artifacts. We recreate a portion of the Stanford NLI corpus using GPT-4, Llama-2 and Mistral 7b, and train hypothesis-only classifiers to determine whether LLM-elicited hypotheses contain annotation artifacts. On our LLM-elicited NLI datasets, BERT-based hypothesis-only classifiers achieve between 86-96% accuracy, indicating these datasets contain hypothesis-only artifacts. We also find frequent "give-aways" in LLM-generated hypotheses, e.g. the phrase "swimming in a pool" appears in more than 10,000 contradictions generated by GPT-4. Our analysis provides empirical evidence that well-attested biases in NLI can persist in LLM-generated data.

Related papers

• CSS: Contrastive Semantic Similarity for Uncertainty Quantification of LLMs [1.515687944002438]
  We propose Contrastive Semantic Similarity, a module to obtain similarity features for measuring uncertainty for text pairs. We conduct extensive experiments with three large language models (LLMs) on several benchmark question-answering datasets. Results show that our proposed method performs better in estimating reliable responses of LLMs than comparable baselines.
  arXiv  Detail & Related papers  (2024-06-05T11:35:44Z)
• Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimension [63.330262740414646]
  We study how to characterize and predict the truthfulness of texts generated from large language models (LLMs) We suggest investigating internal activations and quantifying LLM's truthfulness using the local intrinsic dimension (LID) of model activations.
  arXiv  Detail & Related papers  (2024-02-28T04:56:21Z)
• Native Language Identification with Large Language Models [60.80452362519818]
  We show that GPT models are proficient at NLI classification, with GPT-4 setting a new performance record of 91.7% on the benchmark11 test set in a zero-shot setting. We also show that unlike previous fully-supervised settings, LLMs can perform NLI without being limited to a set of known classes.
  arXiv  Detail & Related papers  (2023-12-13T00:52:15Z)
• Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment [82.60594940370919]
  We propose the FlipFlop experiment to study the multi-turn behavior of Large Language Models (LLMs) We show that models flip their answers on average 46% of the time and that all models see a deterioration of accuracy between their first and final prediction, with an average drop of 17% (the FlipFlop effect) We conduct finetuning experiments on an open-source LLM and find that finetuning on synthetically created data can mitigate - reducing performance deterioration by 60% - but not resolve sycophantic behavior entirely.
  arXiv  Detail & Related papers  (2023-11-14T23:40:22Z)
• Assessing the Reliability of Large Language Model Knowledge [78.38870272050106]
  Large language models (LLMs) have been treated as knowledge bases due to their strong performance in knowledge probing tasks. How do we evaluate the capabilities of LLMs to consistently produce factually correct answers? We propose MOdel kNowledge relIabiliTy scORe (MONITOR), a novel metric designed to directly measure LLMs' factual reliability.
  arXiv  Detail & Related papers  (2023-10-15T12:40:30Z)
• Generating with Confidence: Uncertainty Quantification for Black-box Large Language Models [37.63939774027709]
  Large language models (LLMs) specializing in natural language generation (NLG) have recently started exhibiting promising capabilities. We propose and compare several confidence/uncertainty measures, applying them to *selective NLG* where unreliable results could either be ignored or yielded for further assessment. Results reveal that a simple measure for the semantic dispersion can be a reliable predictor of the quality of LLM responses.
  arXiv  Detail & Related papers  (2023-05-30T16:31:26Z)
• Sources of Hallucination by Large Language Models on Inference Tasks [16.644096408742325]
  Large Language Models (LLMs) are claimed to be capable of Natural Language Inference (NLI) We present a series of behavioral studies on several LLM families which probe their behavior using controlled experiments.
  arXiv  Detail & Related papers  (2023-05-23T22:24:44Z)
• Statistical Knowledge Assessment for Large Language Models [79.07989821512128]
  Given varying prompts regarding a factoid question, can a large language model (LLM) reliably generate factually correct answers? We propose KaRR, a statistical approach to assess factual knowledge for LLMs. Our results reveal that the knowledge in LLMs with the same backbone architecture adheres to the scaling law, while tuning on instruction-following data sometimes compromises the model's capability to generate factually correct text reliably.
  arXiv  Detail & Related papers  (2023-05-17T18:54:37Z)
• The Internal State of an LLM Knows When It's Lying [18.886091925252174]
  Large Language Models (LLMs) have shown exceptional performance in various tasks. One of their most prominent drawbacks is generating inaccurate or false information with a confident tone. We provide evidence that the LLM's internal state can be used to reveal the truthfulness of statements.
  arXiv  Detail & Related papers  (2023-04-26T02:49:38Z)
• Beyond Distributional Hypothesis: Let Language Models Learn Meaning-Text Correspondence [45.9949173746044]
  We show that large-size pre-trained language models (PLMs) do not satisfy the logical negation property (LNP) We propose a novel intermediate training task, names meaning-matching, designed to directly learn a meaning-text correspondence. We find that the task enables PLMs to learn lexical semantic information.
  arXiv  Detail & Related papers  (2022-05-08T08:37:36Z)
• Automatically Identifying Semantic Bias in Crowdsourced Natural Language Inference Datasets [78.6856732729301]
  We introduce a model-driven, unsupervised technique to find "bias clusters" in a learned embedding space of hypotheses in NLI datasets. interventions and additional rounds of labeling can be performed to ameliorate the semantic bias of the hypothesis distribution of a dataset.
  arXiv  Detail & Related papers  (2021-12-16T22:49:01Z)

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.