Search papers, labs, and topics across Lattice.
Affiliation:
6
0
9
8
LLM agents achieve dramatic efficiency gains in long-horizon research by offloading conversational history to a central harness, enabling focused, role-specific contexts for each task.
Even advanced LLMs struggle to replicate human researchers' ability to find inspiring prior work, highlighting a significant gap in AI's scientific intuition.
Regularized Recursive Self-Improvement of Agent Harnesses (RRSI) is introduced, which incorporates the principles of regularizations into harness self-improvement by constraining the evolution candidate proposal and selection.
Harness optimization can significantly enhance model performance, as demonstrated by WHALE's impressive accuracy gains across multiple domains.
SPIRAL achieves a remarkable 15% performance increase by combining sequential, parallel, and aggregative reasoning in language models.
LLM performance hinges on the code around the model, and Meta-Harness proves that automating the design of this "harness" can significantly boost results across diverse tasks.
