AIFIRST Show HN: LLM Attention Visualization
The brief
A Show HN project presents an interactive web tool for visualizing attention patterns inside large language models.
Key points
- Users can inspect how attention is distributed across tokens during inference, providing a hands-on view of LLM internals without running local code.
Sources
- HNishamf.dev