Tag

evaluation

Prism shipped a 27B that fits a laptop. We put it through a battery of agentic traps to see if it reasons like its size.

Parallel decoding of audio tokens in a text-to-speech model

Decoding 16 residual codebooks in one pass instead of sixteen. On-device.

A diffusion model drafts, an autoregressive model verifies. Acceptance measured on eight prompts.