相关论文: Are PPO-ed Language Models Hackable?

Pre-trained transformer language models have shown remarkable performance on a variety of NLP tasks. However, recent research has suggested that phrase-level representations in these models reflect heavy influences of lexical content, but…

Safe Reinforcement Learning (Safe RL) aims to train an RL agent to maximize its performance in real-world environments while adhering to safety constraints, as exceeding safety violation limits can result in severe consequences. In this…

The wave of pre-training language models has been continuously improving the quality of the machine-generated conversations, however, some of the generated responses still suffer from excessive repetition, sometimes repeating words from…

How emotions are expressed depends on the context and domain. On X (formerly Twitter), for instance, an author might simply use the hashtag #anger, while in a news headline, emotions are typically written in a more polite, indirect manner.…

Reinforcement Learning from Verifiable Rewards (RLVR) has recently shown that large language models (LLMs) can develop their own reasoning without direct supervision. However, applications in the medical domain, specifically for question…

Language model (LM) post-training (or alignment) involves maximizing a reward function that is derived from preference annotations. Direct Preference Optimization (DPO) is a popular offline alignment method that trains a policy directly on…

For many applications of reinforcement learning it can be more convenient to specify both a reward function and constraints, rather than trying to design behavior through the reward function. For example, systems that physically interact…

Direct alignment algorithms have proven an effective step for aligning language models to human-desired behaviors. Current variants of the Direct Preference Optimization objective have focused on a strict setting where all tokens are…

A centerpiece of the ever-popular reinforcement learning from human feedback (RLHF) approach to fine-tuning autoregressive language models is the explicit training of a reward model to emulate human feedback, distinct from the language…

Reinforcement learning (RL) can align language models with non-differentiable reward signals, such as human preferences. However, a major challenge arises from the sparsity of these reward signals - typically, there is only a single reward…

Despite advances in AI alignment, large language models (LLMs) remain vulnerable to adversarial attacks or jailbreaking, in which adversaries can modify prompts to induce unwanted behavior. While some defenses have been proposed, they have…

Proximal Policy Optimization (PPO) is a widely used reinforcement learning algorithm that heavily relies on accurate advantage estimates for stable and efficient training. However, raw advantage signals can exhibit significant variance,…

Recent studies have demonstrated the emerging capabilities of foundation models like ChatGPT in several fields, including affective computing. However, accessing these emerging capabilities is facilitated through prompt engineering. Despite…

Expressive generative models have advanced robotic manipulation by capturing complex, multi-modal action distributions over temporally extended trajectories. However, fine-tuning these policies via RL remains challenging due to instability…

Group Relative Policy Optimization (GRPO) has significantly advanced the training of large language models and enhanced their reasoning capabilities, while it remains susceptible to instability due to the use of hard clipping. Soft Adaptive…

Can we construct a neural model that is inductively biased towards learning human languages? Motivated by this question, we aim at constructing an informative prior over neural weights, in order to adapt quickly to held-out languages in the…

Post-training processes are essential phases in grounding pre-trained language models to real-world tasks, with learning from demonstrations or preference signals playing a crucial role in this adaptation. We present a unified theoretical…

This study presents a thorough examination of various Generative Pretrained Transformer (GPT) methodologies in sentiment analysis, specifically in the context of Task 4 on the SemEval 2017 dataset. Three primary strategies are employed: 1)…

Pretrained language models often do not perform tasks in ways that are in line with our preferences, e.g., generating offensive text or factually incorrect summaries. Recent work approaches the above issue by learning from a simple form of…

Reinforcement learning from human feedback (RLHF) is the mainstream paradigm used to align large language models (LLMs) with human preferences. Yet existing RLHF heavily relies on accurate and informative reward models, which are vulnerable…