2 papers
cs.AI2026
Unlocking Noisy Real-World Corpora for Foundation Model Pre-Training via Quality-Aware Tokenization
Arvid E. Gollwitzer, Paridhi Latawa, David de Gruijl +2
Current tokenization methods process sequential data without accounting for signal quality, limiting their effectiveness on noisy real-world corpora. We present QA-Token (Quality-A…
cs.CL2025
Can LLMs Reason Over Non-Text Modalities in a Training-Free Manner? A Case Study with In-Context Representation Learning
Tianle Zhang, Wanlong Fang, Jonathan Woo +3
The remarkable performance of Large Language Models (LLMs) can be enhanced with test-time computation, which relies on external tools and even other deep learning models. However,…