Zed's team details how they built Zeta2, their improved edit prediction model. Key improvements include richer input context (finer-grained edit history, LSP-resolved type/symbol definitions), a switch from Qwen 2.5 Coder (7B) to Seed Coder (8B) as the base model, and a knowledge distillation pipeline using Claude Sonnet as the teacher model. They addressed the 'reversal problem' where the model incorrectly deleted intentional user edits by improving teacher prompting and edit granularity. Training data shifted from synthetic GitHub commit examples to opt-in real user traces from open source repos, yielding ~250-300k training requests per week. The result is a 30% better acceptance rate and faster responses, validated through dogfooding, shadow releases, and gradual rollout.
Table of contents
Knowledge distillationCollecting the right training dataThe reversal problemSwitching the base modelHow to know when it's time to shipWhat's nextIntroducing Zed AIWe Rebuilt Zeta from the Training Data UpChoose Your Edit Prediction Provider4 Comments
Sort: