jevgram heatmap shows which words look AI-written, not just one score
kieranklaassen · x · 2026-09-26
In "What the Jev" #3, Kieran Klaassen highlights jevgram (by Matt Hammer): instead of one AI-detection score, it asks the model the same question about every 1-, 2-, and 3-word window and paints a heatmap pinpointing exactly which phrases trigger the verdict.
The prompt is swappable — ask "does this person sound upset?" and the same sliding-window mechanism becomes an entirely different text-analysis tool.
More from coding & agent
- Building a Claim-Verification Agent Workflow for Research Paper Writing — dupa1234s · 2026-09-26
- Open-source Reef Launches /reefine: Agents Grow New Abilities From a Single Prompt — pliang279 · 2026-09-26
- Google DeepMind launches Polaris team to build eval frameworks for AI coding agents — JMateosGarcia · 2026-09-26
- 28 vibe coding resources actually worth keeping, from design galleries to full build prompts — GCWebDesigner · 2026-09-26
- New AI Coder Meme: Live Long Enough and You End Up Building a Harness — justalexoki · 2026-09-26
- Dev builds isometric game entirely in Claude Code, mixing Seedream and Sunburst art — Dead_Fox_7 · 2026-09-26