KramaBench
PulseAugur coverage of KramaBench — every cluster mentioning KramaBench across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
EMMA system enhances multi-agent workflow synthesis with structured memory
Researchers have introduced EMMA, a novel system for synthesizing multi-agent workflows using structured episodic memory. EMMA addresses limitations in current LLM agent deployment by making update ownership and admissi…
-
New TEAL method evolves reusable tools for coding agents without labels
Researchers have developed TEAL (Tool Evolution for Analytics over data Lakes), a novel unsupervised method for evolving reusable tools for coding agents. TEAL operates without ground-truth labels, using an Explorer age…
-
New SANA framework diagnoses LLM agent failures in data lake QA
Researchers have introduced SANA, a diagnostic framework designed to evaluate the performance of Large Language Model (LLM) agents in exploratory question answering (EQA) over massive data lakes. SANA breaks down end-to…