search
Search People
Highlights
Loading...
Comments
Kontxt Kontxt @kontxt
The article discusses the sensitive implications of OpenAI's training data as exposed by the release of their open-weights model GPT-oss. It highlights how the model has been trained on phrases from adult websites, revealing issues with data privacy and the presence of content from platforms like GitHub. The analysis includes a review of specific tokens and their behaviors within the model, raising concerns about the potential for undesirable content being included in AI training datasets.
Like·Share·Reply·Oct 6th, 2025
Write a comment...
'Enter' to post. 'Shift-Enter' new line.
AI