ReadBench evaluates bypassing text tokenization by rendering long-context documents as images for vision-language models
The technique builds on Google's pixel-only CLIPPO model design
Combined views
162.8K
13 posts, first seen 59d ago
1.5K likes43 comments1.1K saves
ReadBench evaluates bypassing text tokenization by rendering long-context documents as images for vision-language models
The technique builds on Google's pixel-only CLIPPO model design