Search papers, labs, and topics across Lattice.
6
0
7
21
Cross-lingual safety transfer in LLMs is a mirage, with harmful prompts losing over 90% of their safety signals in low-resource languages.
Models trained on AfriSUD reveal a striking syntax gap, highlighting the inadequacy of existing architectures for capturing the complexities of African languages.
Bridging the scientific knowledge gap for hundreds of millions, AfriScience-MT pioneers document-level scientific machine translation for six African languages.
Annotating data in temporal proximity is the single biggest factor in high-quality labels, dwarfing individual annotator speed or tweet content.
Unlearning one demographic group in CLIP doesn't eliminate bias, it just shifts it, primarily along gender lines, revealing a concerning gender-dominant structure in the model's embedding space.
A large-scale, multilingual dataset for online polarization detection is now available, offering a benchmark for future research in this critical area.
