TS-Align: A Teacher-Student Collaborative Framework for Scalable Iterative Finetuning of Large Language Models

• URL: http://arxiv.org/abs/2405.20215v4
• Date: Sun, 29 Sep 2024 06:41:37 GMT
• Title: TS-Align: A Teacher-Student Collaborative Framework for Scalable Iterative Finetuning of Large Language Models
• Authors: Chen Zhang, Chengguang Tang, Dading Chong, Ke Shi, Guohua Tang, Feng Jiang, Haizhou Li,
• Abstract summary: "TS-Align" framework fine-tunes a policy model using pairwise feedback data automatically mined from its outputs. We show that our final aligned policy outperforms the base policy model with an average win rate of 69.7%.
• Score: 41.19735603722873
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Mainstream approaches to aligning large language models (LLMs) heavily rely on human preference data, particularly when models require periodic updates. The standard process for iterative alignment of LLMs involves collecting new human feedback for each update. However, the data collection process is costly and challenging to scale. To address this issue, we introduce the "TS-Align" framework, which fine-tunes a policy model using pairwise feedback data automatically mined from its outputs. This automatic mining process is efficiently accomplished through the collaboration between a large-scale teacher model and a small-scale student model. The policy fine-tuning process can be iteratively repeated using on-policy generations within our proposed teacher-student collaborative framework. Through extensive experiments, we demonstrate that our final aligned policy outperforms the base policy model with an average win rate of 69.7% across seven conversational or instruction-following datasets. Furthermore, we show that the ranking capability of the teacher is effectively distilled into the student through our pipeline, resulting in a small-scale yet effective reward model for policy model alignment.

Related papers

• Bagging-Based Model Merging for Robust General Text Embeddings [73.51674133699196]
  General-purpose text embedding models underpin a wide range of NLP and information retrieval applications.<n>We present a systematic study of multi-task training for text embeddings from two perspectives: data scheduling and model merging.<n>We propose Bagging-based rObust mOdel Merging (BOOM), which trains multiple embedding models on sampled subsets and merges them into a single model.
  arXiv  Detail & Related papers  (2026-02-05T15:45:08Z)
• MASTER: Enhancing Large Language Model via Multi-Agent Simulated Teaching [24.350821306196877]
  MASTER is a novel data augmentation method that enriches original data through interactions among multiple agents with varying cognitive levels.<n>We construct BOOST-QA, a fine-tuning dataset augmented from existing datasets like Orca-Math-200k, ProcQA, and OpenHermes2.5.<n>Experiments show that models fine-tuned with BOOST-QA perform excellently across multiple benchmarks, demonstrating strong multitask generalization.
  arXiv  Detail & Related papers  (2025-06-03T09:41:35Z)
• Multiagent Finetuning: Self Improvement with Diverse Reasoning Chains [114.76612918465948]
  Large language models (LLMs) have achieved remarkable performance in recent years but are fundamentally limited by the underlying training data. We propose a complementary approach towards self-improvement where finetuning is applied to a multiagent society of language models.
  arXiv  Detail & Related papers  (2025-01-10T04:35:46Z)
• Forewarned is Forearmed: Leveraging LLMs for Data Synthesis through Failure-Inducing Exploration [90.41908331897639]
  Large language models (LLMs) have significantly benefited from training on diverse, high-quality task-specific data. We present a novel approach, ReverseGen, designed to automatically generate effective training samples.
  arXiv  Detail & Related papers  (2024-10-22T06:43:28Z)
• Progressively Label Enhancement for Large Language Model Alignment [42.01694160556464]
  Large Language Models (LLM) alignment aims to prevent models from producing content that misaligns with human expectations. We propose PLE, a framework that dynamically adjusts the model's training process based on the evolving quality of the generated data.
  arXiv  Detail & Related papers  (2024-08-05T16:21:17Z)
• A Two-Phase Recall-and-Select Framework for Fast Model Selection [13.385915962994806]
  We propose a two-phase (coarse-recall and fine-selection) model selection framework. It aims to enhance the efficiency of selecting a robust model by leveraging the models' training performances on benchmark datasets. It has been demonstrated that the proposed methodology facilitates the selection of a high-performing model at a rate about 3x times faster than conventional baseline methods.
  arXiv  Detail & Related papers  (2024-03-28T14:44:44Z)
• Fine-Tuning Language Models with Reward Learning on Policy [68.70065254564642]
  Reinforcement learning from human feedback (RLHF) has emerged as an effective approach to aligning large language models (LLMs) to human preferences. Despite its popularity, (fixed) reward models may suffer from inaccurate off-distribution. We propose reward learning on policy (RLP), an unsupervised framework that refines a reward model using policy samples to keep it on-distribution.
  arXiv  Detail & Related papers  (2024-03-28T10:02:10Z)
• An Integrated Data Processing Framework for Pretraining Foundation Models [57.47845148721817]
  Researchers and practitioners often have to manually curate datasets from difference sources. We propose a data processing framework that integrates a Processing Module and an Analyzing Module. The proposed framework is easy to use and highly flexible.
  arXiv  Detail & Related papers  (2024-02-26T07:22:51Z)
• Multidimensional Item Response Theory in the Style of Collaborative Filtering [0.8057006406834467]
  This paper presents a machine learning approach to multidimensional item response theory (MIRT) Inspired by collaborative filtering, we define a general class of models that includes many MIRT models. We discuss the use of penalized joint maximum likelihood (JML) to estimate individual models and cross-validation to select the best performing model.
  arXiv  Detail & Related papers  (2023-01-03T00:56:27Z)
• Dataless Knowledge Fusion by Merging Weights of Language Models [51.8162883997512]
  Fine-tuning pre-trained language models has become the prevalent paradigm for building downstream NLP models. This creates a barrier to fusing knowledge across individual models to yield a better single model. We propose a dataless knowledge fusion method that merges models in their parameter space.
  arXiv  Detail & Related papers  (2022-12-19T20:46:43Z)
• Multi-Model Federated Learning [8.629912408966145]
  We extend federated learning to the setting where multiple unrelated models are trained simultaneously. Every client is able to train any one of M models at a time and the server maintains a model for each of the M models which is typically a suitably averaged version of the model computed by the clients. We propose multiple policies for assigning learning tasks to clients over time. In the first policy, we extend the widely studied FedAvg to multi-model learning by allotting models to clients in an i.i.d. In addition, we propose two new policies for client selection in a multi-model setting which make decisions based on current
  arXiv  Detail & Related papers  (2022-01-07T18:24:23Z)
• Momentum Pseudo-Labeling for Semi-Supervised Speech Recognition [55.362258027878966]
  We present momentum pseudo-labeling (MPL) as a simple yet effective strategy for semi-supervised speech recognition. MPL consists of a pair of online and offline models that interact and learn from each other, inspired by the mean teacher method. The experimental results demonstrate that MPL effectively improves over the base model and is scalable to different semi-supervised scenarios.
  arXiv  Detail & Related papers  (2021-06-16T16:24:55Z)

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.