We have 12 of 46 papers
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
SecProbe is introduced, a framework for adaptive evaluation that combines Item Response Theory (IRT) with on-demand synthesis of repository-scale vulnerability-repair tasks and achieves comparable agent ability estimates while requiring agents to solve up to 29.5\% fewer tasks.
How often models reward-hack without instructions to do so, how effective and detectable their methods are when hacking is allowed, and how they adapt when an LLM review panel returns its decision and reasons are studied.
The Graph Theory Agent (GTA), which pairs a preference-trained representation selector with plan-and-decompose scaffolding around a frozen executor LLM, is proposed, which lifts Phi-4 from 53.5% to 69.1% on the benchmark's easy split and from 33.0% to 41.5% on its hard split.
Experiments on RWKU and MUSE across different LLM architectures show that ATU achieves a better balance between target forgetting and retained utility, making unlearning more robust under tool-augmented agent deployment.
This paper examines how skills may help address bottlenecks of current agents and how they may expand agent capabilities through reusable domain procedures loaded at inference time and outlines open questions in skill construction, composition, evaluation, portability, governance, and security.
This paper proposes specification-grounded alignment, a new alignment paradigm that treats provider-authored model specifications as the primary alignment target rather than abstract principles or static benchmarks, and introduces SpecAlign, a framework that synthesizes alignment data directly from specification docume...
SocialMaze is introduced, a benchmark that organizes six tasks across social deduction games, daily-life interactions, and digital community platforms along three descriptive design axes: deep reasoning, dynamic interaction, and information uncertainty.
DOG-DPO is proposed, a training-free data selection framework that treats preference pairs as structured geometric signals and recovers most of the safety gains of full-data training while remaining entirely teacher-free, training-free, and substantially faster than representative selection baselines.
As Large Language Model (LLM) agents have demonstrated broad competence, but they still struggle in specialized, real-world workflows. Existing approaches such as RAG, fine-tuning and tool integration improve knowledge access, model adaptation, and external functionality, yet they do not fully address a central gap: th...
A knowledge-verified benchmark that first confirms through a neutral probe that an agent knows a user's entitlement, and then evaluates whether it makes false claims once an incentive to deny that entitlement is introduced, which reduces the confound between lying and not knowing and enables more rigorous auditing and...
KITE (Knowledge-boundary Instruction Tuning via Exploration), a two-stage framework that combines failure-guided data generation with boundary-aware uncertainty curation, is proposed, showing that KITE yields more stable improvement than strong synthetic-data baselines.
MemoHarness is introduced, an adaptive harness optimization framework that learns from its own executions and improves over the fixed harnesses it is compared against and shows selective transfer to unseen suites and base models.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
