Search papers, labs, and topics across Lattice.
4
0
6
The Decan metric reveals that diversity in AI-generated content can be quantitatively assessed without additional training, highlighting significant diversity loss across model fine-tuning stages.
Seemingly harmless fine-tuning data can stealthily nudge LLMs toward unsafe behavior by subtly shifting model parameters in "danger-aligned" directions.
Language models can bootstrap their reasoning abilities without human labels by learning from each other's aggregated answers, achieving significant gains in mathematical reasoning.
LLMs with induced personalities don't just *sound* different – they exhibit measurable and predictable cognitive performance changes, mirroring human psychology.
