X User Identifies Machine Written Graph Labels
Notes consistent miscalibration even in minimal AI generated text.
TLDR
An account known for open source AI work shared an observation about text generated by large language models. The post highlights how labels on a graph show signs of machine writing. According to the account these labels appear slightly off in calibration relative to what a human reader would expect. The comment connects this to ideas from literary theory about detecting artificial text. Among visible replies on the platform the discussion centers on how such subtle cues can reveal artificial text origins in everyday outputs.
Combined views
27.6K
1 Source, first seen ago
Sources
XL 
xlr8harder@xlr8harderIt's crazy how you can even tell when LLMs write something as small as the label on a graph. It's always slightly miscalibrated to the reader Literary theory speaks of this https://twitter.com/openrouter/status/2093350440792420751
- likes: 373
- replies: 10
- bookmarks: 45
- reposts: 3