Twitxy Saved

Kamal Maher

Verified account This account is verified. Learn more

@fluorocore

AI x biosecurity @SecureBio @pivotal_org
Previously PhD @MIT BS @Cornell
London kmaherx.github.io Joined June 2016
890 Following 244 Followers 57 Posts

Latest posts 3

Kamal Maher Verified account @fluorocore
Super cool work, and verbalization of soft prompts for interpretability is slept on imo.
Beam search seems to be the authors' critical component for making soft prompts verbalizable.

But I think there may be a smoother-brained way… 🧵
Nathan Hu Verified account @NathanHu12 Sep 17
New paper! In subliminal learning, LLMs transmit traits (e.g. loving cats) though seemingly unrelated data (e.g. numbers). We proactively detect these effects as readable prompts. To do so, we use the surprising ability of models to verbalize learned soft prompts.🧵
View quoted post on X
1 replies 0 reposts 9 likes 1.1K views View on X
Kamal Maher Verified account @fluorocore
Recent literature suggests that unsupervised tissue region identification in spatial omics data requires complex statistical or graph neural network models.

However, in our new preprint, we introduce a much simpler approach: biorxiv.org/content/10.110…
www.biorxiv.org Mitigating autocorrelation during spatially resolved transcriptomics data analysis Several computational methods have recently been developed for characterizing molecular tissue regions in spatially resolved transcriptomics (SRT) data. However, each method fundamentally relies on...