Jev Combined with Legal Workflows
A user suggests combining Jev with Astra could supercharge legal workflows, highlighting its potential in legal compliance.
Jev Performance on CLERC Retrieval Benchmark
On a subset of CLERC queries, Laya without fine-tuning scored 2.5%, below BM25 (5%) and FastLate (10%); Jev reached 30%, Luna 22.5%, and with prompt tuning Luna gets 52.5–55%.
I ran it on a subset of CLERC queries (recovering a missing citation’s source). Laya, without fine-tuning, scored 2.5% top-1, below BM25 (5%) and FastLate (10%). Jev 30%, Luna 22.5% , with bit of prompt tuning Luna gets 52.5–55%,
· 0 likesOpen on X
More like this
Jev completes Traditional Chinese traffic-accident legal analysis in one API call
A user tested TypeSafe AI's Jev on an 11-page AI-written Traditional Chinese traffic-accident analysis, producing 15 typed decisions (5 legal findings × 3 questions) in one API call: 831 ms end-to-end, ~$0.0003 (7,645 input tokens; output free), and 15/15 schema passes.
Building a Legal Chatbot Database Router with Jev
A developer shares a classifier test built with TypeSafe AI's Jev (via OpenRouter) to determine which database a legal chatbot should search.
