Are the Values of LLMs Structurally Aligned with Humans? A Causal Perspective

• URL: http://arxiv.org/abs/2501.00581v2
• Date: Sun, 23 Feb 2025 16:33:30 GMT
• Title: Are the Values of LLMs Structurally Aligned with Humans? A Causal Perspective
• Authors: Yipeng Kang, Junqi Wang, Yexin Li, Mengmeng Wang, Wenming Tu, Quansen Wang, Hengli Li, Tingjun Wu, Xue Feng, Fangwei Zhong, Zilong Zheng,
• Abstract summary: We argue that a latent causal value graph underlies the value dimensions of large language models (LLMs) and that, despite alignment training, this structure remains significantly different from human value systems.<n>We leverage these causal value graphs to guide two lightweight value-steering methods: role-based prompting and sparse autoencoder (SAE) steering.<n>Experiments on Gemma-2B-IT and Llama3-8B-IT demonstrate the effectiveness and controllability of our methods.
• Score: 33.19778298286475
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: As large language models (LLMs) become increasingly integrated into critical applications, aligning their behavior with human values presents significant challenges. Current methods, such as Reinforcement Learning from Human Feedback (RLHF), typically focus on a limited set of coarse-grained values and are resource-intensive. Moreover, the correlations between these values remain implicit, leading to unclear explanations for value-steering outcomes. Our work argues that a latent causal value graph underlies the value dimensions of LLMs and that, despite alignment training, this structure remains significantly different from human value systems. We leverage these causal value graphs to guide two lightweight value-steering methods: role-based prompting and sparse autoencoder (SAE) steering, effectively mitigating unexpected side effects. Furthermore, SAE provides a more fine-grained approach to value steering. Experiments on Gemma-2B-IT and Llama3-8B-IT demonstrate the effectiveness and controllability of our methods.

Related papers

• CLASH: Evaluating Language Models on Judging High-Stakes Dilemmas from Multiple Perspectives [3.7731230532888036]
  CLASH (Character perspective-based LLM Assessments in Situations with High-stakes) is a dataset consisting of 345 high-impact dilemmas along with 3,795 individual perspectives of diverse values. Even the strongest models, such as GPT-4o and Claude-Sonnet, achieve less than 50% accuracy in identifying situations where the decision should be ambivalent.
  arXiv  Detail & Related papers  (2025-04-15T02:54:16Z)
• Generative Psycho-Lexical Approach for Constructing Value Systems in Large Language Models [13.513813405118478]
  Large Language Models (LLMs) have raised concerns regarding their elusive intrinsic values. This study addresses the gap by introducing the Generative Psycho-Lexical Approach (GPLA) We propose a psychologically grounded five-factor value system tailored for LLMs.
  arXiv  Detail & Related papers  (2025-02-04T16:10:55Z)
• Value Compass Leaderboard: A Platform for Fundamental and Validated Evaluation of LLMs Values [76.70893269183684]
  Large Language Models (LLMs) achieve remarkable breakthroughs, aligning their values with humans has become imperative. Existing evaluations focus narrowly on safety risks such as bias and toxicity. Existing benchmarks are prone to data contamination. The pluralistic nature of human values across individuals and cultures is largely ignored in measuring LLMs value alignment.
  arXiv  Detail & Related papers  (2025-01-13T05:53:56Z)
• Exploring Knowledge Boundaries in Large Language Models for Retrieval Judgment [56.87031484108484]
  Large Language Models (LLMs) are increasingly recognized for their practical applications. Retrieval-Augmented Generation (RAG) tackles this challenge and has shown a significant impact on LLMs. By minimizing retrieval requests that yield neutral or harmful results, we can effectively reduce both time and computational costs.
  arXiv  Detail & Related papers  (2024-11-09T15:12:28Z)
• Enhancing Multiple Dimensions of Trustworthiness in LLMs via Sparse Activation Control [44.326363467045496]
  Large Language Models (LLMs) have become a critical area of research in Reinforcement Learning from Human Feedback (RLHF) representation engineering offers a new, training-free approach. This technique leverages semantic features to control the representation of LLM's intermediate hidden states. It is difficult to encode various semantic contents, like honesty and safety, into a singular semantic feature.
  arXiv  Detail & Related papers  (2024-11-04T08:36:03Z)
• Impact of Missing Values in Machine Learning: A Comprehensive Analysis [0.0]
  This paper aims to examine the nuanced impact of missing values on machine learning (ML) models. Our analysis focuses on the challenges posed by missing values, including biased inferences, reduced predictive power, and increased computational burdens. The study employs case studies and real-world examples to illustrate the practical implications of addressing missing values.
  arXiv  Detail & Related papers  (2024-10-10T18:31:44Z)
• SEAL: Systematic Error Analysis for Value ALignment [4.2185937778110825]
  Reinforcement Learning from Human Feedback aims to align language models with human values. Despite its importance, the internal mechanisms of RLHF remain poorly understood. This paper introduces new metrics to evaluate the effectiveness of modeling and aligning human values.
  arXiv  Detail & Related papers  (2024-08-16T18:48:30Z)
• CoMMIT: Coordinated Instruction Tuning for Multimodal Large Language Models [68.64605538559312]
  In this paper, we analyze the MLLM instruction tuning from both theoretical and empirical perspectives. Inspired by our findings, we propose a measurement to quantitatively evaluate the learning balance. In addition, we introduce an auxiliary loss regularization method to promote updating of the generation distribution of MLLMs.
  arXiv  Detail & Related papers  (2024-07-29T23:18:55Z)
• Towards Effective Evaluations and Comparisons for LLM Unlearning Methods [97.2995389188179]
  This paper seeks to refine the evaluation of machine unlearning for large language models. It addresses two key challenges -- the robustness of evaluation metrics and the trade-offs between competing goals.
  arXiv  Detail & Related papers  (2024-06-13T14:41:00Z)
• An Empirical Study of Automated Vulnerability Localization with Large Language Models [21.84971967029474]
  Large Language Models (LLMs) have shown potential in various domains, yet their effectiveness in vulnerability localization remains underexplored. Our investigation encompasses 10+ leading LLMs suitable for code analysis, including ChatGPT and various open-source models. We explore the efficacy of these LLMs using 4 distinct paradigms: zero-shot learning, one-shot learning, discriminative fine-tuning, and generative fine-tuning.
  arXiv  Detail & Related papers  (2024-03-30T08:42:10Z)
• How Can LLM Guide RL? A Value-Based Approach [68.55316627400683]
  Reinforcement learning (RL) has become the de facto standard practice for sequential decision-making problems by improving future acting policies with feedback. Recent developments in large language models (LLMs) have showcased impressive capabilities in language understanding and generation, yet they fall short in exploration and self-improvement capabilities. We develop an algorithm named LINVIT that incorporates LLM guidance as a regularization factor in value-based RL, leading to significant reductions in the amount of data needed for learning.
  arXiv  Detail & Related papers  (2024-02-25T20:07:13Z)
• InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance [56.184255657175335]
  We develop textbfInferAligner, a novel inference-time alignment method that utilizes cross-model guidance for harmlessness alignment. Experimental results show that our method can be very effectively applied to domain-specific models in finance, medicine, and mathematics. It significantly diminishes the Attack Success Rate (ASR) of both harmful instructions and jailbreak attacks, while maintaining almost unchanged performance in downstream tasks.
  arXiv  Detail & Related papers  (2024-01-20T10:41:03Z)
• Beyond Imitation: Leveraging Fine-grained Quality Signals for Alignment [105.34140537748546]
  We propose an improved alignment approach named FIGA. Different from prior methods, we incorporate fine-grained quality signals that are derived by contrasting good and bad responses. Our approach has made two major contributions. Firstly, we curate a refined alignment dataset that pairs initial responses and the corresponding revised ones. Secondly, we devise a new loss function can leverage fine-grained quality signals to instruct the learning of LLMs for alignment.
  arXiv  Detail & Related papers  (2023-11-07T15:36:40Z)
• Interpreting Learned Feedback Patterns in Large Language Models [11.601799960959214]
  We train probes to estimate the feedback signal implicit in the activations of a fine-tuned language model. We compare these estimates to the true feedback, measuring how accurate the LFPs are to the fine-tuning feedback. We validate our probes by comparing the neural features they correlate with positive feedback inputs against the features GPT-4 describes and classifies as related to LFPs.
  arXiv  Detail & Related papers  (2023-10-12T09:36:03Z)
• SALMON: Self-Alignment with Instructable Reward Models [80.83323636730341]
  This paper presents a novel approach, namely SALMON, to align base language models with minimal human supervision. We develop an AI assistant named Dromedary-2 with only 6 exemplars for in-context learning and 31 human-defined principles.
  arXiv  Detail & Related papers  (2023-10-09T17:56:53Z)
• Heterogeneous Value Alignment Evaluation for Large Language Models [91.96728871418]
  Large Language Models (LLMs) have made it crucial to align their values with those of humans. We propose a Heterogeneous Value Alignment Evaluation (HVAE) system to assess the success of aligning LLMs with heterogeneous values.
  arXiv  Detail & Related papers  (2023-05-26T02:34:20Z)
• Towards Accurate Knowledge Transfer via Target-awareness Representation Disentanglement [56.40587594647692]
  We propose a novel transfer learning algorithm, introducing the idea of Target-awareness REpresentation Disentanglement (TRED) TRED disentangles the relevant knowledge with respect to the target task from the original source model and used as a regularizer during fine-tuning the target model. Experiments on various real world datasets show that our method stably improves the standard fine-tuning by more than 2% in average.
  arXiv  Detail & Related papers  (2020-10-16T17:45:08Z)
• Accurate and Robust Feature Importance Estimation under Distribution Shifts [49.58991359544005]
  PRoFILE is a novel feature importance estimation method. We show significant improvements over state-of-the-art approaches, both in terms of fidelity and robustness.
  arXiv  Detail & Related papers  (2020-09-30T05:29:01Z)

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.