On Positional Bias of Faithfulness for Long-form Summarization

• URL: http://arxiv.org/abs/2410.23609v1
• Date: Thu, 31 Oct 2024 03:50:15 GMT
• Title: On Positional Bias of Faithfulness for Long-form Summarization
• Authors: David Wan, Jesse Vig, Mohit Bansal, Shafiq Joty,
• Abstract summary: Large Language Models (LLMs) often exhibit positional bias in long-context settings, under-attending to information in the middle of inputs. We investigate the presence of this bias in long-form summarization, its impact on faithfulness, and various techniques to mitigate this bias.
• Score: 83.63283027830657
• License:
• Abstract: Large Language Models (LLMs) often exhibit positional bias in long-context settings, under-attending to information in the middle of inputs. We investigate the presence of this bias in long-form summarization, its impact on faithfulness, and various techniques to mitigate this bias. To consistently evaluate faithfulness, we first compile a benchmark of eight human-annotated long-form summarization datasets and perform a meta-evaluation of faithfulness metrics. We show that LLM-based faithfulness metrics, though effective with full-context inputs, remain sensitive to document order, indicating positional bias. Analyzing LLM-generated summaries across six datasets, we find a "U-shaped" trend in faithfulness, where LLMs faithfully summarize the beginning and end of documents but neglect middle content. Perturbing document order similarly reveals models are less faithful when important documents are placed in the middle of the input. We find that this behavior is partly due to shifting focus with context length: as context increases, summaries become less faithful, but beyond a certain length, faithfulness improves as the model focuses on the end. Finally, we experiment with different generation techniques to reduce positional bias and find that prompting techniques effectively direct model attention to specific positions, whereas more sophisticated approaches offer limited improvements. Our data and code are available in https://github.com/meetdavidwan/longformfact.

Related papers

• Distance between Relevant Information Pieces Causes Bias in Long-Context LLMs [50.40165119718928]
  LongPiBench is a benchmark designed to assess positional bias involving multiple pieces of relevant information. These experiments reveal that while most current models are robust against the "lost in the middle" issue, there exist significant biases related to the spacing of relevant information pieces.
  arXiv  Detail & Related papers  (2024-10-18T17:41:19Z)
• Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilization [97.84156490765457]
  Large language models (LLMs) struggle to capture relevant information located in the middle of their input. This phenomenon has been known as the lost-in-the-middle problem. We show found-in-the-middle achieves better performance in locating relevant information within a long context.
  arXiv  Detail & Related papers  (2024-06-23T04:35:42Z)
• FABLES: Evaluating faithfulness and content selection in book-length summarization [55.50680057160788]
  In this paper, we conduct the first large-scale human evaluation of faithfulness and content selection on book-length documents. We collect FABLES, a dataset of annotations on 3,158 claims made in LLM-generated summaries of 26 books, at a cost of $5.2K USD. An analysis of the annotations reveals that most unfaithful claims relate to events and character states, and they generally require indirect reasoning over the narrative to invalidate.
  arXiv  Detail & Related papers  (2024-04-01T17:33:38Z)
• On Context Utilization in Summarization with Large Language Models [83.84459732796302]
  Large language models (LLMs) excel in abstractive summarization tasks, delivering fluent and pertinent summaries. Recent advancements have extended their capabilities to handle long-input contexts, exceeding 100k tokens. We conduct the first comprehensive study on context utilization and position bias in summarization.
  arXiv  Detail & Related papers  (2023-10-16T16:45:12Z)
• Bias in News Summarization: Measures, Pitfalls and Corpora [4.917075909999548]
  We introduce definitions for biased behaviours in summarization models, along with practical operationalizations. We measure gender bias in English summaries generated by both purpose-built summarization models and general purpose chat models. We find content selection in single document summarization to be largely unaffected by gender bias, while hallucinations exhibit evidence of bias.
  arXiv  Detail & Related papers  (2023-09-14T22:20:27Z)
• Evaluation of Faithfulness Using the Longest Supported Subsequence [52.27522262537075]
  We introduce a novel approach to evaluate faithfulness of machine-generated text by computing the longest noncontinuous of the claim that is supported by the context. Using a new human-annotated dataset, we finetune a model to generate Longest Supported Subsequence (LSS) Our proposed metric demonstrates an 18% enhancement over the prevailing state-of-the-art metric for faithfulness on our dataset.
  arXiv  Detail & Related papers  (2023-08-23T14:18:44Z)
• LongEval: Guidelines for Human Evaluation of Faithfulness in Long-form Summarization [48.02158981582502]
  LongEval is a set of guidelines for human evaluation of faithfulness in long-form summaries. We deploy LongEval in annotation studies on two long-form summarization datasets in different domains.
  arXiv  Detail & Related papers  (2023-01-30T21:31:48Z)

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.