One invisible space makes Claude Code and other LLM graders flip verdicts
maverick_man1111 · reddit · 2026-10-28
A developer testing an open-source eval project found Claude Code marked correct Python code as wrong and vice versa — caused by a single invisible Unicode character (NBSP).
The same flaw exists in LangChain OpenEvals, Promptfoo, and Nvidia SkillEvaluator. The fix: auto-generate a note listing the exact Unicode characters and positions without altering the original answer, which drastically improved judgments. He reproduced the issue in Claude Code's native plugin eval grader too.
He submitted PRs to Nvidia, OpenEvals, and Promptfoo; Nvidia's is already approved. Full report at driftproofhq.com/invisible-space.
Related event: Invisible Unicode space flips AI grading results in code evals(2 posts)→
More from coding & agent
- Sentient OS launches on-device AI that reads your digital life nightly with a local LLM — rohanpaul_ai · 2026-10-28
- DeepMind: AI agents spot bad advice 92% of the time—and follow it anyway — alex_verem · 2026-10-28
- Delta brings real-time multiplayer collaboration to Claude Code via a share link — xiaohu · 2026-10-28
- JetBrains Junie Local ships for RTX Spark and Windows, runs coding agent token-free — kimmonismus · 2026-10-28
- Metatron: a System 1 controller that lets coding agents command humanoid robots — ChongZzZhang · 2026-10-28
- Two days tuning one Tesla P40: 35B MoE with full 262k context at ~60 t/s — sargetun123 · 2026-10-28