Preference-Guided Reflective Sampling for Aligning Language Models

• URL: http://arxiv.org/abs/2408.12163v2
• Date: Fri, 4 Oct 2024 11:40:58 GMT
• Title: Preference-Guided Reflective Sampling for Aligning Language Models
• Authors: Hai Ye, Hwee Tou Ng,
• Abstract summary: Iterative data generation and model re-training can effectively align large language models(LLMs) to human preferences. In this work, we propose Preference-Guided Reflective Sampling (PRS) Unlike random sampling, PRS employs a tree-based generation framework to enable more efficient sampling. PRS shows strong performance when applied in iterative offline RL training.
• Score: 27.69410513313001
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Iterative data generation and model re-training can effectively align large language models(LLMs) to human preferences. The process of data sampling is crucial, as it significantly influences the success of policy improvement. Repeated random sampling is a widely used method that independently queries the model multiple times to generate outputs. In this work, we propose a more effective sampling method, named Preference-Guided Reflective Sampling (PRS). Unlike random sampling, PRS employs a tree-based generation framework to enable more efficient sampling. It leverages adaptive self-refinement techniques to better explore the sampling space. By specifying user preferences in natural language, PRS can further optimize response generation according to these preferences. As a result, PRS can align models to diverse user preferences. Our experiments demonstrate that PRS generates higher-quality responses with significantly higher rewards. On AlpacaEval and Arena-Hard, PRS substantially outperforms repeated random sampling in best-of-$N$ sampling. Moreover, PRS shows strong performance when applied in iterative offline RL training.

Related papers

• Finding the Sweet Spot: Preference Data Construction for Scaling Preference Optimization [66.67988187816185]
  We aim to emphscale up the number of on-policy samples via repeated random sampling to improve alignment performance. Our experiments reveal that this strategy leads to a emphdecline in performance as the sample size increases. We introduce a scalable preference data construction strategy that consistently enhances model performance as the sample scale increases.
  arXiv  Detail & Related papers  (2025-02-24T04:22:57Z)
• Clear Preferences Leave Traces: Reference Model-Guided Sampling for Preference Learning [59.11519451499754]
  Direct Preference Optimization (DPO) has emerged as a de-facto approach for aligning language models with human preferences. Recent work has shown DPO's effectiveness relies on training data quality. We discover that reference model probability space naturally detects high-quality training samples.
  arXiv  Detail & Related papers  (2025-01-25T07:21:50Z)
• Preference Optimization with Multi-Sample Comparisons [53.02717574375549]
  We introduce a novel approach that extends post-training to include multi-sample comparisons. These approaches fail to capture critical characteristics such as generative diversity and bias. We demonstrate that multi-sample comparison is more effective in optimizing collective characteristics than single-sample comparison.
  arXiv  Detail & Related papers  (2024-10-16T00:59:19Z)
• Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation [51.127054971591924]
  We introduce a new generative self-evaluation scheme designed to adaptively reduce the number of generated samples. We demonstrate that 74% of the improvement from using 16 samples can be achieved with only 1.2 samples on average.
  arXiv  Detail & Related papers  (2024-10-03T17:47:29Z)
• Balancing Diversity and Risk in LLM Sampling: How to Select Your Method and Parameter for Open-Ended Text Generation [60.493180081319785]
  We propose a systematic way to estimate the intrinsic capacity of a truncation sampling method by considering the trade-off between diversity and risk at each decoding step. Our work provides a comprehensive comparison between existing truncation sampling methods, as well as their recommended parameters as a guideline for users.
  arXiv  Detail & Related papers  (2024-08-24T14:14:32Z)
• Turning Up the Heat: Min-p Sampling for Creative and Coherent LLM Outputs [4.122612309805664]
  Large Language Models (LLMs) generate text by sampling the next token from a probability distribution over the vocabulary at each decoding step. We propose min-p sampling, a dynamic truncation method that adjusts the sampling threshold based on the model's confidence by scaling according to the top token's probability. We conduct extensive experiments on benchmarks including GPQA, GSM8K, and AlpacaEval Creative Writing, demonstrating that min-p sampling improves both the quality and diversity of generated text, particularly at high temperatures.
  arXiv  Detail & Related papers  (2024-07-01T08:37:25Z)
• 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)
• RS-DPO: A Hybrid Rejection Sampling and Direct Preference Optimization Method for Alignment of Large Language Models [7.676477609461592]
  Reinforcement learning from human feedback (RLHF) has been extensively employed to align large language models with user intent. DPO relies on contrastive responses generated from human annotator and alternative LLM, instead of the policy model. In this paper, we address both challenges by systematically combining sampling rejection (RS) and DPO. Our proposed method effectively fine-tunes LLMs with limited resource environments, leading to improved alignment with user intent.
  arXiv  Detail & Related papers  (2024-02-15T16:00:58Z)
• AutoSAM: Towards Automatic Sampling of User Behaviors for Sequential Recommender Systems [48.461157194277504]
  We propose a general automatic sampling framework, named AutoSAM, to non-uniformly treat historical behaviors. Specifically, AutoSAM augments the standard sequential recommendation architecture with an additional sampler layer to adaptively learn the skew distribution of the raw input. We theoretically design multi-objective sampling rewards including Future Prediction and Sequence Perplexity, and then optimize the whole framework in an end-to-end manner.
  arXiv  Detail & Related papers  (2023-11-01T09:25:21Z)
• Sampling Through the Lens of Sequential Decision Making [9.101505546901999]
  We propose a reward-guided sampling strategy called Adaptive Sample with Reward (ASR) Our approach optimally adjusts the sampling process to achieve optimal performance. Empirical results in information retrieval and clustering demonstrate ASR's superb performance across different datasets.
  arXiv  Detail & Related papers  (2022-08-17T04:01:29Z)
• Reparameterized Sampling for Generative Adversarial Networks [71.30132908130581]
  We propose REP-GAN, a novel sampling method that allows general dependent proposals by REizing the Markov chains into the latent space of the generator. Empirically, extensive experiments on synthetic and real datasets demonstrate that our REP-GAN largely improves the sample efficiency and obtains better sample quality simultaneously.
  arXiv  Detail & Related papers  (2021-07-01T10:34:55Z)
• Local policy search with Bayesian optimization [73.0364959221845]
  Reinforcement learning aims to find an optimal policy by interaction with an environment. Policy gradients for local search are often obtained from random perturbations. We develop an algorithm utilizing a probabilistic model of the objective function and its gradient.
  arXiv  Detail & Related papers  (2021-06-22T16:07:02Z)
• Reinforced Data Sampling for Model Diversification [15.547681142342846]
  This paper proposes a new Reinforced Data Sampling (RDS) method to learn how to sample data adequately. We formulate the optimisation problem of model diversification $delta-div$ in data sampling to maximise learning potentials and optimum allocation by injecting model diversity. Our results suggest that the trainable sampling for model diversification is useful for competition organisers, researchers, or even starters to pursue full potentials of various machine learning tasks.
  arXiv  Detail & Related papers  (2020-06-12T11:46:13Z)

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.