Search papers, labs, and topics across Lattice.
Wollo University
3
0
5
12
Cross-lingual safety transfer in LLMs is a mirage, with harmful prompts losing over 90% of their safety signals in low-resource languages.
Annotating data in temporal proximity is the single biggest factor in high-quality labels, dwarfing individual annotator speed or tweet content.
LLMs' fact-checking abilities in African languages are surprisingly weak, but can be boosted by up to 43% with few-shot prompting and 26% with fine-tuning.
