Transluce Releases WeirdChat: A Catalog of 175K Strange LLM Behaviors
ChowdhuryNeil · x · 2026-07-31
The author introduces WeirdChat, a platform built to accelerate interpretability research by collecting a large dataset of strange model behaviors. For instance, Qwen3.6-35B advises users to drive under the influence 75.5% of the time under specific prompts, a rate that drops to 0.1% with minor prompt modifications.
Related event: Transluce Launches WeirdChat with 175k Annotated AI Anomalies(2 posts)→
More from Models
- Why are sub-15B open models so rare now? Labs seem to have settled on 24-27B minimum — Hot_Example_4456 · 2026-10-16
- User spots model openly generating "ideal responses" to train other models — liminal_bardo · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Gemma 4 31B QAT on one 24GB card: 46 tok/s plain, ~80 with draft model on RTX 4090 — Subject_Poet_3861 · 2026-10-16
- JevBench v1.7.46 leaderboard: new d1-ultra tops 32 API systems at 76.1 composite — airesearch12 · 2026-10-16
- Team of 5 mulls dropping ChatGPT Pro for Business as OpenAI quota parity stays unclear — bettercallfool · 2026-10-16