Arian Hosseini

Putting the Value Back in RL: Better Test-Time Scaling by Unifying LLM Reasoners With Verifiers

Morgane M Moss

2025-05-07

ArXiv (preprint)

Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models

Shengyi Huang

The dominant paradigm for RLHF is online and on-policy RL: synchronously generating from the large language model (LLM) policy, labelling wi… (see more)th a reward model, and learning using feedback on the LLM's own outputs. While performant, this paradigm is computationally inefficient. Inspired by classical deep RL literature, we propose separating generation and learning in RLHF. This enables asynchronous generation of new samples while simultaneously training on old samples, leading to faster training and more compute-optimal scaling. However, asynchronous training relies on an underexplored regime, online but off-policy RLHF: learning on samples from previous iterations of our model. To understand the challenges in this regime, we investigate a fundamental question: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? Among several RLHF algorithms we tested, we find that online DPO is most robust to off-policy data, and robustness increases with the scale of the policy model. We study further compute optimizations for asynchronous RLHF but find that they come at a performance cost, giving rise to a trade-off. Finally, we verify the scalability of asynchronous RLHF by training LLaMA 3.1 8B on an instruction-following task 40% faster than a synchronous run while matching final performance.

2025-01-22

ICLR.cc/2025/Conference (poster)

Faster, More Efficient RLHF through Off-Policy Asynchronous Learning

Shengyi Huang

To achieve state-of-the-art chatbots, large language models are finetuned with reinforcement learning (RL), frequently to optimize human fee… (see more)dback (RLHF). This process is computationally expensive and can take weeks. Offline approaches, like DPO, learn on a static dataset and are efficient but not performant. The dominant paradigm, online and on-policy---synchronously generating from the model, labelling with a reward model, and learning on feedback from the model's own outputs---is performant but not efficient. Following prior work in the generall deep RL setting, we propose separating the actor and learner in RLHF. This enables the asynchronously generation of new samples while learning on prior samples, thus leading to overall faster training and better scaling. But this requires a novel regime for RLHF, online but off-policy: learning on samples from a previous version of our model. We ask a fundamental question: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? We find that a contrastive loss, Online DPO, is most robust to off-policy data and that robustness increases with the scale of the policy model. We show even further compute optimizations but demonstrate that they come at a performance cost, giving rise to a trade-off. Finally, we verify our design choices by training LLaMA 3.1 8B with RLHF as a helpful chatbot in half the time of a synchronous run while matching final performance.

2025-01-22

ICLR.cc/2025/Conference (poster)

Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models

Shengyi Huang

The dominant paradigm for RLHF is online and on-policy RL: synchronously generating from the large language model (LLM) policy, labelling wi… (see more)th a reward model, and learning using feedback on the LLM's own outputs. While performant, this paradigm is computationally inefficient. Inspired by classical deep RL literature, we propose separating generation and learning in RLHF. This enables asynchronous generation of new samples while simultaneously training on old samples, leading to faster training and more compute-optimal scaling. However, asynchronous training relies on an underexplored regime, online but off-policy RLHF: learning on samples from previous iterations of our model. To understand the challenges in this regime, we investigate a fundamental question: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? Among several RLHF algorithms we tested, we find that online DPO is most robust to off-policy data, and robustness increases with the scale of the policy model. We study further compute optimizations for asynchronous RLHF but find that they come at a performance cost, giving rise to a trade-off. Finally, we verify the scalability of asynchronous RLHF by training LLaMA 3.1 8B on an instruction-following task 40% faster than a synchronous run while matching final performance.

2024-10-23

ArXiv (preprint)

Faster, More Efficient RLHF through Off-Policy Asynchronous Learning

Shengyi Huang

To achieve state-of-the-art chatbots, large language models are finetuned with reinforcement learning (RL), frequently to optimize human fee… (see more)dback (RLHF). This process is computationally expensive and can take weeks. Offline approaches, like DPO, learn on a static dataset and are efficient but not performant. The dominant paradigm, online and on-policy---synchronously generating from the model, labelling with a reward model, and learning on feedback from the model's own outputs---is performant but not efficient. Following prior work in the generall deep RL setting, we propose separating the actor and learner in RLHF. This enables the asynchronously generation of new samples while learning on prior samples, thus leading to overall faster training and better scaling. But this requires a novel regime for RLHF, online but off-policy: learning on samples from a previous version of our model. We ask a fundamental question: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? We find that a contrastive loss, Online DPO, is most robust to off-policy data and that robustness increases with the scale of the policy model. We show even further compute optimizations but demonstrate that they come at a performance cost, giving rise to a trade-off. Finally, we verify our design choices by training LLaMA 3.1 8B with RLHF as a helpful chatbot in half the time of a synchronous run while matching final performance.

2024-10-10

NeurIPS.cc/2024/Workshop/FITML (poster)

Not All LLM Reasoners Are Created Equal

Daniel Toyama

2024-10-09

NeurIPS.cc/2024/Workshop/Sys2-Reasoning (poster)

Not All LLM Reasoners Are Created Equal

Daniel Toyama

We study the depth of grade-school math (GSM) problem-solving capabilities of LLMs. To this end, we evaluate their performance on pairs of e… (see more)xisting math word problems together so that the answer to the second problem depends on correctly answering the first problem. Our findings reveal a significant reasoning gap in most LLMs, that is performance difference between solving the compositional pairs and solving each question independently. This gap is more pronounced in smaller, more cost-efficient, and math-specialized models. Moreover, instruction-tuning recipes and code generation have varying effects across LLM sizes, while finetuning on GSM can lead to task overfitting. Our analysis indicates that large reasoning gaps are not because of test-set leakage, but due to distraction from additional context and poor second-hop reasoning. Overall, LLMs exhibit systematic differences in their reasoning abilities, despite what their performance on standard benchmarks indicates.

2024-10-02

ArXiv (preprint)

V-STaR: Training Verifiers for Self-Taught Reasoners

Xingdi Yuan

Common self-improvement approaches for large language models (LLMs), such as STaR (Zelikman et al., 2022), iteratively fine-tune LLMs on sel… (see more)f-generated solutions to improve their problem-solving ability. However, these approaches discard the large amounts of incorrect solutions generated during this process, potentially neglecting valuable information in such solutions. To address this shortcoming, we propose V-STaR that utilizes both the correct and incorrect solutions generated during the self-improvement process to train a verifier using DPO that judges correctness of model-generated solutions. This verifier is used at inference time to select one solution among many candidate solutions. Running V-STaR for multiple iterations results in progressively better reasoners and verifiers, delivering a 4% to 17% test accuracy improvement over existing self-improvement and verification approaches on common code generation and math reasoning benchmarks with LLaMA2 models.

2024-07-10

colmweb.org/COLM/2024/Conference (accepted)

V-STaR: Training Verifiers for Self-Taught Reasoners

Xingdi Yuan

Common self-improvement approaches for large language models (LLMs), such as STaR, iteratively fine-tune LLMs on self-generated solutions to… (see more) improve their problem-solving ability. However, these approaches discard the large amounts of incorrect solutions generated during this process, potentially neglecting valuable information in such solutions. To address this shortcoming, we propose V-STaR that utilizes both the correct and incorrect solutions generated during the self-improvement process to train a verifier using DPO that judges correctness of model-generated solutions. This verifier is used at inference time to select one solution among many candidate solutions. Running V-STaR for multiple iterations results in progressively better reasoners and verifiers, delivering a 4% to 17% test accuracy improvement over existing self-improvement and verification approaches on common code generation and math reasoning benchmarks with LLaMA2 models.

2024-02-09

ArXiv (preprint)

V-STaR: Training Verifiers for Self-Taught Reasoners

Xingdi Yuan

Common self-improvement approaches for large language models (LLMs), such as STaR, iteratively fine-tune LLMs on self-generated solutions to… (see more) improve their problem-solving ability. However, these approaches discard the large amounts of incorrect solutions generated during this process, potentially neglecting valuable information in such solutions. To address this shortcoming, we propose V-STaR that utilizes both the correct and incorrect solutions generated during the self-improvement process to train a verifier using DPO that judges correctness of model-generated solutions. This verifier is used at inference time to select one solution among many candidate solutions. Running V-STaR for multiple iterations results in progressively better reasoners and verifiers, delivering a 4% to 17% test accuracy improvement over existing self-improvement and verification approaches on common code generation and math reasoning benchmarks with LLaMA2 models.

2024-02-09

ArXiv (preprint)

V-STaR: Training Verifiers for Self-Taught Reasoners

Xingdi Yuan

Common self-improvement approaches for large language models (LLMs), such as STaR (Zelikman et al., 2022), iteratively fine-tune LLMs on sel… (see more)f-generated solutions to improve their problem-solving ability. However, these approaches discard the large amounts of incorrect solutions generated during this process, potentially neglecting valuable information in such solutions. To address this shortcoming, we propose V-STaR that utilizes both the correct and incorrect solutions generated during the self-improvement process to train a verifier using DPO that judges correctness of model-generated solutions. This verifier is used at inference time to select one solution among many candidate solutions. Running V-STaR for multiple iterations results in progressively better reasoners and verifiers, delivering a 4% to 17% test accuracy improvement over existing self-improvement and verification approaches on common code generation and math reasoning benchmarks with LLaMA2 models.

2024-02-09

ArXiv (preprint)