Why logits can simulate a jev-style classifier, explained
michellechen · x · 2026-09-19
The author explains the mechanism: while System One models' internals are unknown, their interface can be approximated with any LLM. Models emit logits for every vocabulary token; softmax turns them into probabilities and logprobs express them in log space — carrying more information than the finally selected tokens. With the right inputs and prompting, these scores yield a probability distribution that simulates a jev-style classification model.
Related event: Developers Turn LLM Logprobs into Probabilistic Decision Outputs(2 posts)→
More from coding & agent
- macOS 27 ships with mlx_whisper built in, letting agents transcribe video locally — vista8 · 2026-09-19
- Custom /jev command in OpenCode turns LLM rambling into instant yes/no answers — TheMoonMidas · 2026-09-19
- Dev pipelines Codex output to Muse via Google Drive, calls it insane — RachelVT42 · 2026-09-19
- Building a Self-Improving Content Skill with Claude Code Artifacts and MCP — alexgoughcooper · 2026-09-19
- Jevable site catalogs 342 demos of Typeface's Jev, from $0.0039 flight search to intent-scored spreadsheets — gaganghotra_ · 2026-09-19
- Omar Khattab: AI-era programs still need real languages, declarative specs and DSPy-style modules — lateinteraction · 2026-09-19