NOVER: Incentive Training for Language Models via Verifier-Free Reinforcement Learning

TL;DR

NOVER introduces a verifier-free reinforcement learning framework for language models that enhances reasoning capabilities without external verifiers, using only supervised fine-tuning data, and outperforms some large reasoning models.

Contribution

NOVER presents a novel incentive training method that eliminates the need for external verifiers, broadening applicability and improving performance of language models.

Findings

Outperforms distilled models from large reasoning models by 7.7%.

Enables incentive training across diverse text-to-text tasks.

Supports inverse incentive training for further optimization.

Abstract

Recent advances such as DeepSeek R1-Zero highlight the effectiveness of incentive training, a reinforcement learning paradigm that computes rewards solely based on the final answer part of a language model's output, thereby encouraging the generation of intermediate reasoning steps. However, these methods fundamentally rely on external verifiers, which limits their applicability to domains like mathematics and coding where such verifiers are readily available. Although reward models can serve as verifiers, they require high-quality annotated data and are costly to train. In this work, we propose NOVER, NO-VERifier Reinforcement Learning, a general reinforcement learning framework that requires only standard supervised fine-tuning data with no need for an external verifier. NOVER enables incentive training across a wide range of text-to-text tasks and outperforms the model of the same size distilled from large reasoning models such as DeepSeek R1 671B by 7.7 percent. Moreover, the flexibility of NOVER enables new possibilities for optimizing large language models, such as inverse incentive training.