Related papers: Reward Hacking Benchmark: Measuring Exploits in LL…

Recent neural theorem provers use reinforcement learning with verifiable rewards (RLVR), where proof assistants provide binary correctness signals. While verifiable rewards are cheap and scalable without reward hacking issues, they suffer…

As LLM-based agents increasingly rely on external tools, it is important to evaluate their ability to sustain tool-grounded reasoning beyond familiar workflows and short-range interactions. We introduce AgentEscapeBench, an…

We propose ContextRL, a novel framework that leverages context augmentation to overcome these bottlenecks. Specifically, to enhance Identifiability, we provide the reward model with full reference solutions as context, enabling fine-grained…

An appropriate reward function is of paramount importance in specifying a task in reinforcement learning (RL). Yet, it is known to be extremely challenging in practice to design a correct reward function for even simple tasks.…

Recent advances in frontier large language models have enabled code review agents that operate in open-ended, reasoning-intensive settings. However, the lack of standardized benchmarks and granular evaluation protocols makes it difficult to…

Reinforcement Learning with Verifiable Rewards (RLVR) is an emerging paradigm that significantly boosts a Large Language Model's (LLM's) reasoning abilities on complex logical tasks, such as mathematics and programming. However, we…

Text-to-SQL is a challenging task involving multiple reasoning-intensive subtasks, including natural language understanding, database schema comprehension, and precise SQL query formulation. Existing approaches often rely on handcrafted…

LLM agents are increasingly relevant to research domains such as vulnerability discovery. Yet, the strongest systems remain closed and cloud-only, making them resource-intensive, difficult to reproduce, and unsuitable for work involving…

Human-designed reward functions for reinforcement learning (RL) agents are frequently misaligned with the humans' true, unobservable objectives, and thus act only as proxies. Optimizing for a misspecified proxy reward function often induces…

Large language model (LLM) agents are moving beyond prompting alone. ChatGPT marked the rise of general-purpose LLM assistants, DeepSeek showed that on-policy reinforcement learning with verifiable rewards can improve reasoning and tool…

LLM-based agents score well on search benchmarks, yet real users consistently find results unsatisfying, revealing a persistent evaluation-experience gap. We attribute this gap to existing benchmarks' reliance on over-specified queries,…

Large Language Model (LLM) agents can leverage multiple turns and tools to solve complex tasks, with prompt-based approaches achieving strong performance. This work demonstrates that Reinforcement Learning (RL) can push capabilities…

The current state-of-the-art backdoor attacks against Reinforcement Learning (RL) rely upon unrealistically permissive access models, that assume the attacker can read (or even write) the victim's policy parameters, observations, or…

While numerous works have focused on devising efficient algorithms for reinforcement learning (RL) with uniformly bounded rewards, it remains an open question whether sample or time-efficient algorithms for RL with large state-action space…

Semantic textual similartiy (STS) and information retrieval tasks (IR) tasks have been the two major avenues to record the progress of embedding models in the past few years. Under the emerging Retrieval-augmented Generation (RAG) paradigm,…

Agent skills, which are reusable, domain-specific knowledge artifacts, have become a popular mechanism for extending LLM-based agents, yet formally benchmarking skill usage performance remains scarce. Existing skill benchmarking efforts…

In offline reinforcement learning (RL) agents are trained using a logged dataset. It appears to be the most natural route to attack real-life applications because in domains such as healthcare and robotics interactions with the environment…

Existing works increasingly adopt memory-centric mechanisms to process long contexts in a segment manner, and effective memory management is one of the key capabilities that enables large language models to effectively propagate information…

Reinforcement Learning with Human Feedback (RLHF) and Reinforcement Learning with Verifiable Rewards (RLVR) are the main RL paradigms used in LLM post-training, each offering distinct advantages. However, RLHF struggles with…

Large language models (LLMs) are increasingly deployed as autonomous agents in offensive cybersecurity. In this paper, we reveal an interesting phenomenon: different agents exhibit distinct attack patterns. Specifically, each agent exhibits…