Every processed story in chronological order, with the newest coverage first. Filter by tag, source, or score to drill in.
MemToolAgent demonstrates that structured memory management — without any LLM fine-tuning — can substantially improve tool-use accuracy, with an 80% relative gain on NESTFUL showing the approach's potential to close the gap between static LLM agents and agents that learn from experience.
The results show that structured verifier feedback — not just more data — can unlock large performance gains for LLM agents on formal reasoning tasks, pointing toward a concrete path for verifier-guided program synthesis.
AutoPDE's explicit strategy representation closes a key gap in LLM-based PDE solvers, where numerical decisions previously remained hidden in code and were difficult to inspect or correct when solves failed.
The benchmark exposes a large performance gap between current frontier LLM agents and human-level proficiency on standardized Office tasks, demonstrating that fine-grained document automation remains a significant unsolved challenge despite recent advances in code generation.
This survey provides a unified, systems-oriented framework for a rapidly expanding but fragmented field, identifying both the dominant attack surfaces and the gaps in current defenses and benchmarks that leave deployed LLM agents exposed.
The paper addresses a core limitation of existing LLM agent memory systems — difficulty with evidence aggregation and fact revision across sessions — by introducing a structured, maintainable architecture that improves both how memory is organized and how it is retrieved.
HIPIF directly targets long-context interference — a problem existing hierarchical RL and credit-assignment methods leave unaddressed — by folding completed subgoal histories, offering a path to more reliable LLM agent performance on extended, multi-turn tasks.
WebChallenger demonstrates that near-frontier web agent performance is achievable with open-weight models at a fraction of the inference cost of proprietary reasoning systems, by addressing architectural gaps rather than scaling model size.
MIRAGE demonstrates that covert encoding by LLM agents — which evades output-side detection — leaves a consistent internal signature that can be monitored in real time, substantially improving detection accuracy over surface-level approaches.
The paper provides the first operational definition of "agent harness" with a shared vocabulary, enabling consistent engineering practice and scientific comparison of agentic coding systems.