Laura Ruis (@lauraruis) 's Twitter Profile
Laura Ruis

@lauraruis

PhD supervised by @_rockt and @egrefen, also part time @cohere. Language and LLMs. Spent time at FAIR, Google, and NYU (@LakeBrenden). She/her.

ID: 1186333456457973760

linkhttp://lauraruis.github.io calendar_today21-10-2019 17:32:03

1,1K Tweet

4,4K Takipçi

711 Takip Edilen

Laura Ruis (@lauraruis) 's Twitter Profile Photo
Laura Ruis
@lauraruis

a month ago

LLMs can be programmed by backprop 🔎 In our new preprint, we show they can act as fuzzy program interpreters and databases. After being ‘programmed’ with next-token prediction, they can retrieve, evaluate, and even *compose* programs at test time, without seeing I/O examples.

LLMs can be programmed by backprop 🔎 In our new preprint, we show they can act as fuzzy program interpreters and databases. After being ‘programmed’ with next-token prediction, they can retrieve, evaluate, and even *compose* programs at test time, without seeing I/O examples.