Related papers: When Is Enough Not Enough? Illusory Completion in …

In continual RL we want agents capable of never-ending learning, and yet our evaluation methodologies do not reflect this. The standard practice in RL is to assume unfettered access to the deployment environment for the full lifetime of the…

Large Language Model (LLM)-based agentic systems have shown growing promise in tackling complex, multi-step tasks through autonomous planning, reasoning, and interaction with external environments. However, the stochastic nature of LLM…

Run the same LLM agent on the same task twice: do you get the same behavior? We find the answer is often no. In a study of 3,000 agent runs across three models (Llama 3.1 70B, GPT-4o, and Claude Sonnet 4.5) on HotpotQA, we observe that…

Multimodal large language models (MLLMs) are increasingly deployed in open-ended, real-world environments where inputs are messy, underspecified, and not always trustworthy. Unlike curated benchmarks, these settings frequently involve…

Many real-world questions appear deceptively simple yet implicitly demand two capabilities: (i) systematic coverage of a bounded knowledge universe and (ii) compositional set-based reasoning over that universe, a phenomenon we term "the tip…

Deducing whodunit proves challenging for LLM agents. In this paper, we implement a text-based multi-agent version of the classic board game Clue as a rule-based testbed for evaluating multi-step deductive reasoning, with six agents drawn…

Large language models are increasingly deployed as specialized agents that plan, call tools, and take actions over extended horizons. Yet many existing evaluations assume a "clean interface" where dynamics are specified and stable, tools…

In a conversation, a helpful assistant must reliably follow user directives, even as they refine, modify, or contradict earlier requests. Yet most instruction-following benchmarks focus on single-turn or short multi-turn scenarios, leaving…

Large language models (LLMs) often fail to scale their performance on long-context tasks performance in line with the context lengths they support. This gap is commonly attributed to retrieval failures -- the models' inability to identify…

Recent studies show that the reasoning capabilities of Large Language Models (LLMs) can be improved by applying Reinforcement Learning (RL) to question-answering (QA) tasks in areas such as math and coding. With a long context length, LLMs…

Large language models (LLMs) are increasingly integrated into high-stakes decision-making. Inspired by the theory of \emph{inattentional blindness} in human cognition, we investigate whether LLMs, trained on human-preferred corpora that…

Large Language Models (LLMs) excel at many tasks but still struggle with a critical ability for LLM-based agents: asking good questions for resolving ambiguity in user requests. While prior work has explored information-seeking behavior…

Alternating-time Temporal Logic (ATL*) is a central logic for multiagent systems. Its extension to the imperfect information setting (ATL*i ) is well known to have an undecidable model-checking problem when agents have perfect recall.…

LLM-powered search agents are increasingly being used for multi-step information seeking tasks, yet the IR community lacks empirical understanding of how agentic search sessions unfold and how retrieved evidence is reflected in later…

LLMs demonstrate strong performance on code benchmarks, yet consistent reasoning across forward and backward execution remains elusive. We present RoundTripCodeEval (RTCE), a benchmark of four code execution reasoning tasks that evaluates…

The evaluation of Deep Research Agents is a critical challenge, as conventional outcome-based metrics fail to capture the nuances of their complex reasoning. Current evaluation faces two primary challenges: 1) a reliance on singular metrics…

Existing benchmarks for Large Language Model (LLM) agents focus on task completion under idealistic settings but overlook reliability in real-world, user-facing applications. In domains, such as in-car voice assistants, users often issue…

Large reasoning models (LRMs) have shown remarkable progress on complex reasoning tasks. However, some questions posed to LRMs are inherently unanswerable, such as math problems lacking sufficient conditions. We find that LRMs continually…

Outcome-rewarded Large Language Models (LLMs) have demonstrated remarkable success in mathematical problem-solving. However, this success often masks a critical issue: models frequently achieve correct answers through fundamentally unsound…

Deep search agents, which autonomously iterate through multi-turn web-based reasoning, represent a promising paradigm for complex information-seeking tasks. However, current agents suffer from critical inefficiency: they conduct excessive…