Elvis Saravia says structured data works well for LLM-based agents, especially in complex codebases, and that context graphs may have broader uses.
He points to Harvey's report, where context graphs raised LAB task accuracy: Firm Knowledge rose from 65% to 80%, precedent lists from 67% to 80%, and matter counting from 53% to 83%. Harvey also reports a 48% cost reduction for agents given a context graph.
LLM-based agents love structured data.
I've seen this work really well, but mostly in complex codebases. But I think context graphs have broader applications, as shown in this Harvey report on legal knowledge tasks.
We are working on this at @dair_ai for research-related tasks. More to share soon.
We built context graphs to help agents work through long-horizon legal tasks. Starting with Legal Agent Bench tasks in 11 categories like M&A diligence and contract negotiation, we had agents build context graphs of the underlying documents (without access to the eval tasks or grading rubrics). Context graphs dramatically improved agent performance on LAB tasks: - Firm Knowledge tasks: 65% → 80% - Precedent lists tasks: 67% → 80% - Matter counting tasks: 53% → 83% They also boost agent efficiency: we found a 48% cost reduction for agents given access to a context graph to assist their work. Full video below - more to come soon.View quoted post on X
Source: elvis · x.comPublished · added here
