Jev Text Generation

Text generation with a model that never generates text: jev-gpt asks Jev one typed question per choice — word type, WordNet category, a group of about 200 words split by embeddings, then the word — about 400 calls and 2 cents per prompt.

A set of experiments that uses TypeSafe's Jev as a decoder through its Choice primitive, sampling ASCII characters, BPE tokens and grammar-constrained SQL from Jev's probability distributions, and measuring next-character loss against uniform sampling on tiny-shakespeare.

A chatbot with no language model: every word is picked from roughly 4,700 options by Jev Choice questions, with confidence-driven lookahead, beam search across sentence directions and a self-critique loop that rewrites weak sentences. Code for the paper and the live demo at talktojev.com.