Exploring Safety-Utility Trade-Offs in Personalized Language Models

• URL: http://arxiv.org/abs/2406.11107v1
• Date: Mon, 17 Jun 2024 00:17:11 GMT
• Title: Exploring Safety-Utility Trade-Offs in Personalized Language Models
• Authors: Anvesh Rao Vijjini, Somnath Basu Roy Chowdhury, Snigdha Chaturvedi,
• Abstract summary: We show that large language models (LLMs) suffer from personalization bias, where their performance is impacted when they are personalized to a user's identity. We quantify personalization bias by evaluating the performance of LLMs along two axes - safety and utility. We discuss several strategies to mitigate personalization bias using preference tuning and prompt-based defenses.
• Score: 26.792174008353008
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: As large language models (LLMs) become increasingly integrated into daily applications, it is essential to ensure they operate fairly across diverse user demographics. In this work, we show that LLMs suffer from personalization bias, where their performance is impacted when they are personalized to a user's identity. We quantify personalization bias by evaluating the performance of LLMs along two axes - safety and utility. We measure safety by examining how benign LLM responses are to unsafe prompts with and without personalization. We measure utility by evaluating the LLM's performance on various tasks, including general knowledge, mathematical abilities, programming, and reasoning skills. We find that various LLMs, ranging from open-source models like Llama (Touvron et al., 2023) and Mistral (Jiang et al., 2023) to API-based ones like GPT-3.5 and GPT-4o (Ouyang et al., 2022), exhibit significant variance in performance in terms of safety-utility trade-offs depending on the user's identity. Finally, we discuss several strategies to mitigate personalization bias using preference tuning and prompt-based defenses.

Related papers

• SG-Bench: Evaluating LLM Safety Generalization Across Diverse Tasks and Prompt Types [21.683010095703832]
  We develop a novel benchmark to assess the generalization of large language model (LLM) safety across various tasks and prompt types. This benchmark integrates both generative and discriminative evaluation tasks and includes extended data to examine the impact of prompt engineering and jailbreak on LLM safety. Our assessment reveals that most LLMs perform worse on discriminative tasks than generative ones, and are highly susceptible to prompts, indicating poor generalization in safety alignment.
  arXiv  Detail & Related papers  (2024-10-29T11:47:01Z)
• Personalization of Large Language Models: A Survey [131.00650432814268]
  Personalization of Large Language Models (LLMs) has recently become increasingly important with a wide range of applications. Most existing works on personalized LLMs have focused either entirely on (a) personalized text generation or (b) leveraging LLMs for personalization-related downstream applications, such as recommendation systems. We introduce a taxonomy for personalized LLM usage and summarizing the key differences and challenges.
  arXiv  Detail & Related papers  (2024-10-29T04:01:11Z)
• MetaAlign: Align Large Language Models with Diverse Preferences during Inference Time [50.41806216615488]
  Large Language Models (LLMs) acquire extensive knowledge and remarkable abilities from extensive text corpora. To make LLMs more usable, aligning them with human preferences is essential. We propose an effective method, textbf MetaAlign, which aims to help LLMs dynamically align with various explicit or implicit preferences specified at inference time.
  arXiv  Detail & Related papers  (2024-10-18T05:31:13Z)
• Embodied Agent Interface: Benchmarking LLMs for Embodied Decision Making [85.24399869971236]
  We aim to evaluate Large Language Models (LLMs) for embodied decision making. Existing evaluations tend to rely solely on a final success rate. We propose a generalized interface (Embodied Agent Interface) that supports the formalization of various types of tasks.
  arXiv  Detail & Related papers  (2024-10-09T17:59:00Z)
• Justice or Prejudice? Quantifying Biases in LLM-as-a-Judge [84.34545223897578]
  Despite their excellence in many domains, potential issues are under-explored, undermining their reliability and the scope of their utility. We identify 12 key potential biases and propose a new automated bias quantification framework-CALM- which quantifies and analyzes each type of bias in LLM-as-a-Judge. Our work highlights the need for stakeholders to address these issues and remind users to exercise caution in LLM-as-a-Judge applications.
  arXiv  Detail & Related papers  (2024-10-03T17:53:30Z)
• Few-shot Personalization of LLMs with Mis-aligned Responses [40.0349773257245]
  This paper proposes a new approach for a few-shot personalization of large language models (LLMs) Our key idea is to learn a set of personalized prompts for each user by progressively improving the prompts using LLMs. During an iterative process of prompt improvement, we incorporate the contexts of mis-aligned responses by LLMs.
  arXiv  Detail & Related papers  (2024-06-26T18:29:12Z)
• Knowledge Graph Tuning: Real-time Large Language Model Personalization based on Human Feedback [5.778012023739487]
  We propose Knowledge Graph Tuning (KGT) to personalize large language models (LLMs) KGT extracts personalized factual knowledge triples from users' queries and feedback and optimize KGs without modifying the LLM parameters. Experiments with state-of-the-art LLMs, including GPT-2, Llama2, and Llama3, show that KGT significantly improves personalization performance while reducing latency and GPU memory costs.
  arXiv  Detail & Related papers  (2024-05-30T04:57:03Z)
• Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models [52.98743860365194]
  We propose a new fine-tuning method called Self-Play fIne-tuNing (SPIN) At the heart of SPIN lies a self-play mechanism, where the LLM refines its capability by playing against instances of itself. This sheds light on the promise of self-play, enabling the achievement of human-level performance in LLMs without the need for expert opponents.
  arXiv  Detail & Related papers  (2024-01-02T18:53:13Z)
• Supervised Knowledge Makes Large Language Models Better In-context Learners [94.89301696512776]
  Large Language Models (LLMs) exhibit emerging in-context learning abilities through prompt engineering. The challenge of improving the generalizability and factuality of LLMs in natural language understanding and question answering remains under-explored. We propose a framework that enhances the reliability of LLMs as it: 1) generalizes out-of-distribution data, 2) elucidates how LLMs benefit from discriminative models, and 3) minimizes hallucinations in generative tasks.
  arXiv  Detail & Related papers  (2023-12-26T07:24:46Z)
• Do LLMs Understand User Preferences? Evaluating LLMs On User Rating Prediction [15.793007223588672]
  Large Language Models (LLMs) have demonstrated exceptional capabilities in generalizing to new tasks in a zero-shot or few-shot manner. We investigate various LLMs in different sizes, ranging from 250M to 540B parameters and evaluate their performance in zero-shot, few-shot, and fine-tuning scenarios.
  arXiv  Detail & Related papers  (2023-05-10T21:43:42Z)
• Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback [11.895749982167375]
  Large language models (LLMs) are used to generate content for a wide range of tasks, and are set to reach a growing audience in coming years. This intensifies the need to ensure that models are aligned with human preferences and do not produce unsafe, inaccurate or toxic outputs. Personalising LLMs through micro-level preference learning processes may result in models that are better aligned with each user.
  arXiv  Detail & Related papers  (2023-03-09T17:52:07Z)

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.