Related papers: Library Drift: Diagnosing and Fixing a Silent Fail…

Ensuring the safety of Large Language Models (LLMs) is critical for real-world deployment. However, current safety measures often fail to address implicit, domain-specific risks. To investigate this gap, we introduce a dataset of 3,000…

The exponential increase in software vulnerabilities has created an urgent need for automatic vulnerability repair (AVR) solutions. Recent research has formulated AVR as a sequence generation problem and has leveraged large language models…

Reinforcement learning (RL) has been widely used to train LLM agents for multi-turn interactive tasks, but its sample efficiency is severely limited by sparse rewards and long horizons. On-policy self-distillation (OPSD) alleviates this by…

On-policy self-distillation (SD) improves LLM reasoning by using teacher-side privileged information (PI) to turn sparse verifier outcomes into dense token-level supervision. Existing methods usually assume trusted PI, such as reference…

A key aspect of automating predictive machine learning entails the capability of properly triggering the update of the trained model. To this aim, suitable automatic solutions to self-assess the prediction quality and the data distribution…

A central paradox in fine-tuning Large Language Models (LLMs) with Reinforcement Learning with Verifiable Reward (RLVR) is the frequent degradation of multi-attempt performance (Pass@k) despite improvements in single-attempt accuracy…

Reinforcement learning (RL) has enabled complex reasoning abilities in large language models (LLMs). However, most RL algorithms suffer from performance saturation, preventing continued gains as RL training scales. This problem can be…

Large language model (LLM) ecosystems such as Claude Code and ChatGPT increasingly rely on skills: packages of natural-language instructions and executable tools. Once in the LLM's context, skill content cannot be reliably separated from…

An ongoing challenge in neural information processing is: how do neurons adjust their connectivity to improve task performance over time (i.e., actualize learning)? It is widely believed that there is a consistent, synaptic-level learning…

Agent skills today are static artifact: authored once -- by human curation or one-shot generation from parametric knowledge -- and then consumed unchanged, with no mechanism to improve from real use. We propose \textbf{SkillEvolver}, a…

Machine learning (ML) models frequently experience performance degradation when deployed in new contexts. Such degradation is rarely uniform: some subgroups may suffer large performance decay while others may not. Understanding where and…

Machine learning and deep learning-based decision making has become part of today's software. The goal of this work is to ensure that machine learning and deep learning-based systems are as trusted as traditional software. Traditional…

Recent advances in Reinforcement Learning (RL) have underscored its potential for incentivizing reasoning capabilities of Large Language Models (LLMs). However, existing step-level efforts suffer from costly annotations that limit domain…

In statistical modelling the biggest threat is concept drift which makes the model gradually showing deteriorating performance over time. There are state of the art methodologies to detect the impact of concept drift, however general…

Adaptive reward design for deep reinforcement learning (DRL) in multi-beam LEO satellite scheduling is motivated by the intuition that regime-aware reward weights should outperform static ones. We systematically test this intuition and…

As language models (LMs) are increasingly deployed as autonomous agents, their robust adherence to human-assigned objectives becomes crucial for safe operation. When these agents operate independently for extended periods without human…

The alignment of Large Language Models (LLMs) for complex reasoning heavily relies on Reinforcement Learning with Verifiable Rewards (RLVR). However, standard algorithms like GRPO apply sequence-level rewards uniformly to all tokens,…

This study reveals a critical safety blind spot in modern LLMs: learning-style queries, which closely resemble ordinary educational questions, can reliably elicit harmful responses. The learning-style queries are constructed by a novel…

While current software agents powered by large language models (LLMs) and agentic reinforcement learning (RL) can boost programmer productivity, their training data (e.g., GitHub issues and pull requests) and environments (e.g.,…

A trained ML model is deployed on another `test' dataset where target feature values (labels) are unknown. Drift is distribution change between the training and deployment data, which is concerning if model performance changes. For a…