Related papers: Safety Alignment as Continual Learning: Mitigating…

The success of Deepseek-R1 has drawn the LLM community's attention to reinforcement learning (RL) methods like GRPO. However, such rule-based 0/1 outcome reward methods lack the capability to regulate the intermediate reasoning processes…

We study post-training linear autoregressive models with outcome and process rewards. Given a context $\boldsymbol{x}$, the model must predict the response $\boldsymbol{y} \in Y^N$, a sequence of length $N$ that satisfies a $\gamma$ margin…

Recursive self-improvement is moving from theory to practice: modern systems can critique, revise, and evaluate their own outputs, yet iterative self-modification risks subtle alignment drift. We introduce SAHOO, a practical framework to…

In speech deepfake detection (SDD), data augmentation (DA) is commonly used to improve model generalization across varied speech conditions and spoofing attacks. However, during training, the backpropagated gradients from original and…

Reinforcement learning (RL) commonly relies on scalar rewards with limited ability to express temporal, conditional, or safety-critical goals, and can lead to reward hacking. Temporal logic expressible via the more general class of…

LLM post-training typically propagates task gradients through the full depth of the model. Although this end-to-end structure is simple and general, it couples task adaptation to full-depth activation storage, long-range backward…

Safety alignment is crucial to ensure that large language models (LLMs) behave in ways that align with human preferences and prevent harmful actions during inference. However, recent studies show that the alignment can be easily compromised…

Large Language Models (LLMs) are increasingly adopted in high-stakes scenarios, yet their safety mechanisms often remain fragile. Simple jailbreak prompts or even benign fine-tuning can bypass these protocols, underscoring the need to…

While most AI alignment research focuses on preventing models from generating explicitly harmful content, a more subtle risk is emerging: capability-oriented training induced exploitation. We investigate whether language models, when…

We study the common continual learning setup where an overparameterized model is sequentially fitted to a set of jointly realizable tasks. We analyze forgetting, defined as the loss on previously seen tasks, after $k$ iterations. For…

While the wide adoption of refusal training in large language models (LLMs) has showcased improvements in model safety, recent works have highlighted shortcomings due to the shallow nature of these alignment methods. To this end, the work…

Balancing performance trade-off on long-tail (LT) data distributions remains a long-standing challenge. In this paper, we posit that this dilemma stems from a phenomenon called "tail performance degradation" (the model tends to severely…

Large reasoning models (LRMs) achieve strong performance on complex reasoning tasks but often generate harmful responses to malicious user queries. This paper investigates the underlying cause of these safety risks and shows that the issue…

Large language model (LLM) safety classifiers such as Llama Guard are effective at detecting overtly harmful prompts but remain vulnerable to adversarial jailbreak attacks that disguise malicious intent through role-play scenarios,…

Large Language Models (LLMs) rely on safety alignment to produce socially acceptable responses. However, this behavior is known to be brittle: further fine-tuning, even on benign or lightly contaminated data, can degrade safety and…

Recent research has observed that in machine learning optimization, gradient descent (GD) often operates at the edge of stability (EoS) [Cohen, et al., 2021], where the stepsizes are set to be large, resulting in non-monotonic losses…

The post-training of LLMs, which typically consists of the supervised fine-tuning (SFT) stage and the preference learning stage (RLHF or DPO), is crucial to effective and safe LLM applications. The widely adopted approach in post-training…

Stochastic gradient-based descent (SGD), have long been central to training large language models (LLMs). However, their effectiveness is increasingly being questioned, particularly in large-scale applications where empirical evidence…

Multi-Task Learning (MTL) combined with Low-Rank Adaptation (LoRA) has emerged as a promising direction for parameter-efficient deployment of Large Language Models (LLMs). By sharing a single adapter across multiple tasks, one can…

Masked diffusion language models (MDLMs) have emerged as a promising alternative to dominant autoregressive approaches. Although they achieve competitive performance on several tasks, a substantial gap remains in open-ended text generation.…