Search papers, labs, and topics across Lattice.
Palo Alto Networks
2
0
4
A new fingerprinting method for AI skills achieves 77x compression while maintaining per-component identity, enabling nuanced recognition of skill families and tampering.
Forget complex training schemes – pinpointing and tweaking just 20 neurons can flip an LLM from sycophantic to truthful, thanks to a new "perturbation probing" technique.
