Visual pretraining on rendered document pages beats text-only continued pretraining on scientific reasoning benchmarks across multiple model backbones, using about a quarter of the tokens.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Scalable Visual Pretraining for Language Intelligence
Visual pretraining on rendered document pages beats text-only continued pretraining on scientific reasoning benchmarks across multiple model backbones, using about a quarter of the tokens.