dos_desvíos

el deber de dar forma al caos no cesa de aumentar

[dos_desvíos @dosdesvios]

@rasbt Great explanation, stupid question: why cant you just store all the outputs and compare? Too much data?

[El Gato y La Caja @ElGatoyLaCaja]

Leé Hermosas criaturas horrendas, una nota escrita por @dosdesvios, e ilustrada por Nono Pautasso https://t.co/pOgmr40awb

[El Gato y La Caja @ElGatoyLaCaja]

¿Cómo hace Claude para aprender cosas? 🧵

[El Gato y La Caja @ElGatoyLaCaja]

Leé Hermosas criaturas horrendas, escrita por @dosdesvios e ilustrada por Nono Pautasso. https://t.co/y0mt7H4xO9

[dos_desvíos @dosdesvios]

@simonw But how can you tell if it is production-quality code if you don't understand it?

I guess the challenge right now is being more productive than before while keeping control over what you ship, but it's not obvious where the sweet spot is

(Thx for answering btw, love u)

[dos_desvíos @dosdesvios]

@geoffreylitt Ok but there is a point where it would have been faster if you had done all the coding yourself, keeping your understanding in the first place

[dos_desvíos @dosdesvios]

@jxmnop People saying they dont write well... I do think they are way above the avg human its just that people recognize the style and start to reject it. But every writer has a style

[dos_desvíos @dosdesvios]

@StephenLCasper @GoodfireAI "In the past few years, several papers have demoed this kind of thing" would you providr some examples?

[dos_desvíos @dosdesvios]

Esto de los LLM se nos fue absolutamente de las manos

[dos_desvíos @dosdesvios]

Si las black box más black box de la historia se terminan explicando a sí mismas en lenguaje natural va a ser un giro totalmente poético

[Owain Evans @OwainEvans_UK]

New paper:
We train Activation Oracles: LLMs that decode their own neural activations and answer questions about them in natural language.
We find surprising generalization. For instance, our AOs uncover misaligned goals in fine-tuned models, without training to do so.

[dos_desvíos @dosdesvios]

Un efecto colateral MUY positivo del research en interpretability es la cantidad de material didáctico de buena calidad que generó

[dos_desvíos @dosdesvios]

@nickhjiang Thx for ur answer! For that purpose, I could use LDA or any other topic modeling technique, can't I?

[dos_desvíos @dosdesvios]

@ChrisGPotts Love these videos! Please keep uploading practice runs!

[dos_desvíos @dosdesvios]

Cursor es el pináculo de la civilización.

[dos_desvíos @dosdesvios]

@gptcrosa Jaajja atroden crosa

[dos_desvíos @dosdesvios]

@stanfordnlp @tomchen0 Interesting! Will you record it?

[dos_desvíos @dosdesvios]

@simonw This would explain why they usually don't come up with deep or new relations, the same way an encyclopedia stores a lot of knowledge but isn't able to rearrange it. They lack the big picture

[dos_desvíos @dosdesvios]

La historia del NLP puede rastrearse en las notas a las sucesivas ediciones de esta biblia hermosa

[dos_desvíos @dosdesvios]

Relegaron naive bayes al apéndice :(

[Dan Jurafsky @jurafsky]

Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: https://t.co/pjfZsYxSk0