#llm-agents
13 briefs tagged #llm-agents.
Researchers propose aDSL for agentic 3D creation
The system pairs an agent-centric DSL with a multi-agent loop for more robust 3D program generation.
Researchers push structured memory for AI agents
Six new papers propose schema, graph and skill-based systems for more reliable long-term agent memory.
Researchers target memory failures in long-term LLM agents
New papers propose graph, segment and interference methods for maintaining agent memory.
Researchers target LLM exploration beyond temperature
DORA, 3PO and RISE-RL propose new ways to improve exploration in LLM agents and reinforcement learning.
Paper bounds GPU opportunity in LLM-agent control
The study models when agent control paths expose enough concurrent work for GPU execution.
New agent frameworks optimize design tasks with LLM search
AUTO, AutoDesign and SynAct apply iterative LLM agents to code, posters and chip synthesis.
Researchers flag backdoor risks in LLM-agent skills
New papers probe how reusable agent skills can be poisoned, compressed and evolved.
HarnessOpt-Bench tests LLMs on agent harness optimization
The benchmark measures how LLMs improve prompts, tools, memory and orchestration around agents.
New papers target scalable training for LLM agents
EnvACE and State2State reduce reliance on manual environments, expert trajectories and task design.
Researchers target prompt injection in LLM agents
Two papers propose adaptive defenses and automated red-teaming for agent security.
Researchers target prompt injection risks in LLM agents
New papers test attacks and defenses for agents, including robotic systems and indirect prompt injection.
Researchers target sparse rewards in agentic RL
New arXiv papers propose self-distillation methods for denser credit assignment in LLM agents.
New papers test self-distillation for LLM reinforcement learning
Studies propose ICE and OCSD while warning that privileged-information teachers can fail on harder tasks.