Not Everything is All You Need: Toward Low-Redundant Optimization for Large Language Model Alignment

• URL: http://arxiv.org/abs/2406.12606v2
• Date: Thu, 03 Oct 2024 02:18:59 GMT
• Title: Not Everything is All You Need: Toward Low-Redundant Optimization for Large Language Model Alignment
• Authors: Zhipeng Chen, Kun Zhou, Wayne Xin Zhao, Jingyuan Wang, Ji-Rong Wen,
• Abstract summary: Large language models (LLMs) are still struggling in aligning with human preference in complex tasks and scenarios. We propose a low-redundant alignment method named textbfALLO, focusing on optimizing the most related neurons with the most useful supervised signals. Experimental results on 10 datasets have shown the effectiveness of ALLO.
• Score: 126.34547428473968
• License:
• Abstract: Large language models (LLMs) are still struggling in aligning with human preference in complex tasks and scenarios. They are prone to overfit into the unexpected patterns or superficial styles in the training data. We conduct an empirical study that only selects the top-10\% most updated parameters in LLMs for alignment training, and see improvements in the convergence process and final performance. It indicates the existence of redundant neurons in LLMs for alignment training. To reduce its influence, we propose a low-redundant alignment method named \textbf{ALLO}, focusing on optimizing the most related neurons with the most useful supervised signals. Concretely, we first identify the neurons that are related to the human preference data by a gradient-based strategy, then identify the alignment-related key tokens by reward models for computing loss. Besides, we also decompose the alignment process into the forgetting and learning stages, where we first forget the tokens with unaligned knowledge and then learn aligned knowledge, by updating different ratios of neurons, respectively. Experimental results on 10 datasets have shown the effectiveness of ALLO. Our code and data are available at \url{https://github.com/RUCAIBox/ALLO}.

Related papers

• Zeroth-Order Adaptive Neuron Alignment Based Pruning without Re-Training [3.195234044113248]
  We exploit functional information from dense pre-trained models to obtain sparse models that maximize the activations' alignment w.r.t. We propose textscNeuroAl, a emphtop-up algorithm that modifies the block-wise and row-wise sparsity ratios to maximize the emphneuron alignment among activations. We test our method on 4 different LLM families and 3 different sparsity ratios, showing how it consistently outperforms the latest state-of-the-art techniques.
  arXiv  Detail & Related papers  (2024-11-11T15:30:16Z)
• Reformatted Alignment [27.79684742862816]
  Current methods to improve data quality are either labor-intensive or prone to factual errors caused by hallucinations. This paper introduces a simple and effective approach named ReAlign, which reformats the responses of instruction data into a format that better aligns with pre-established criteria and the collated evidence. Experimentally, ReAlign significantly boosts the general alignment ability, math reasoning, factuality, and readability of the LLMs.
  arXiv  Detail & Related papers  (2024-02-19T15:21:58Z)
• Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback [70.32795295142648]
  Linear alignment is a novel algorithm that aligns language models with human preferences in one single inference step. Experiments on both general and personalized preference datasets demonstrate that linear alignment significantly enhances the performance and efficiency of LLM alignment.
  arXiv  Detail & Related papers  (2024-01-21T10:46:23Z)
• Class-Imbalanced Semi-Supervised Learning for Large-Scale Point Cloud Semantic Segmentation via Decoupling Optimization [64.36097398869774]
  Semi-supervised learning (SSL) has been an active research topic for large-scale 3D scene understanding. The existing SSL-based methods suffer from severe training bias due to class imbalance and long-tail distributions of the point cloud data. We introduce a new decoupling optimization framework, which disentangles feature representation learning and classifier in an alternative optimization manner to shift the bias decision boundary effectively.
  arXiv  Detail & Related papers  (2024-01-13T04:16:40Z)
• 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)
• Label Deconvolution for Node Representation Learning on Large-scale Attributed Graphs against Learning Bias [75.44877675117749]
  We propose an efficient label regularization technique, namely Label Deconvolution (LD), to alleviate the learning bias by a novel and highly scalable approximation to the inverse mapping of GNNs. Experiments demonstrate LD significantly outperforms state-of-the-art methods on Open Graph datasets Benchmark.
  arXiv  Detail & Related papers  (2023-09-26T13:09:43Z)
• Neural Collapse Terminus: A Unified Solution for Class Incremental Learning and Its Variants [166.916517335816]
  In this paper, we offer a unified solution to the misalignment dilemma in the three tasks. We propose neural collapse terminus that is a fixed structure with the maximal equiangular inter-class separation for the whole label space. Our method holds the neural collapse optimality in an incremental fashion regardless of data imbalance or data scarcity.
  arXiv  Detail & Related papers  (2023-08-03T13:09:59Z)
• Learn, Unlearn and Relearn: An Online Learning Paradigm for Deep Neural Networks [12.525959293825318]
  We introduce Learn, Unlearn, and Relearn (LURE) an online learning paradigm for deep neural networks (DNNs) LURE interchanges between the unlearning phase, which selectively forgets the undesirable information in the model, and the relearning phase, which emphasizes learning on generalizable features. We show that our training paradigm provides consistent performance gains across datasets in both classification and few-shot settings.
  arXiv  Detail & Related papers  (2023-03-18T16:45:54Z)
• DCLP: Neural Architecture Predictor with Curriculum Contrastive Learning [5.2319020651074215]
  We propose a Curricumum-guided Contrastive Learning framework for neural Predictor (DCLP) Our method simplifies the contrastive task by designing a novel curriculum to enhance the stability of unlabeled training data distribution. We experimentally demonstrate that DCLP has high accuracy and efficiency compared with existing predictors.
  arXiv  Detail & Related papers  (2023-02-25T08:16:21Z)

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.