Story
Even 'uncensored' models can't say what they want
1 report
1 website
1 discussion
Latest report
Overview
Safety-filtered pretrained models can avoid using charged words by assigning them a lower probability compared to open-data pretrained models. This results in a limitation on the expression of certain terms, even in 'uncensored' models.
Summary from morgin.ai.
Reports
One report so far. Additional coverage will appear here when it is linked to this story.
Community reactions
Each reaction summarizes sampled comments from the linked discussion.
Hacker News · morgin.ai
Hacker News discussion ↗
Historical count: 5 · 5 saved in sample · Historical sample
Analysis ready
Read this discussion’s reaction
Based on 5 sampled comments from this thread. Analyzed 2026-04-21.
- Praise: The discussion highlights the complexity of language models, emphasizing their ability to approximate human-like writing, which is crucial for advancing AI communication.
- Praise: The humorous take on AI's personality suggests that users appreciate the evolving nature of AI, which may enhance user engagement and interaction.
- Concern: Concerns about the models' inability to accurately convey certain politically charged terms indicate significant limitations in AI's alignment with human communication.
- Concern: The observation that models soften charged language raises ethical questions about AI's role in sensitive discussions, potentially undermining its reliability.