Maja Trebacz @majatrebacz Twitter profile

Last Seen Profiles

@anlas7031

@dlw_t8684

@MrBigPop

@BaronLeSanj

@xxvraii_

@Tante_Binal69

@PKYouthClub

@fkhmf581

@Ningndng

@aespakorn

@kthomas2_

@K_Nakamura

@Nurgulcetak

@yngyng246132

@AlshqyqAtl28877

@kic_suna

@EbrahimBaeshen

@4u5nvo4mddnv

@aragon1500

@loluwa_o

@MukhafiAhmad

@yngyng246132

@bokeplokalmalam

@sungvids

@jandakembangstw

@OlivReigh

@bnt265042266250

@PantlessTheo

@Mahmoud66602603

@Gia9154

@ivyiswylder

@wn_olp

@Abo70004a

@AlshqyqAtl28877

@stw_pdg

@stwmaniax

Maja Trebacz

@majatrebacz

3 months

Btw, I joined @OpenAI and this is what I’ve been up to so far. We've just released a paper on training LLM critics to enhance human feedback for training LLMs. Kudos to the incredible team — @nmca , @gadzin1203 , @agentydragon , Juan, @janleike Excited for what's ahead 🚀

OpenAI

@OpenAI

3 months

We’ve trained a model, CriticGPT, to catch bugs in GPT-4’s code. We’re starting to integrate such models into our RLHF alignment pipeline to help humans supervise AI on difficult tasks:

611

595

Maja Trebacz

@majatrebacz

2 years

𓅪𓅪𓅪 Sparrow 𓅪𓅪𓅪 It was amazing to work on this dialogue agent and train it from human feedback! Sparrow searches Google to improve and back up its claims, and follows a set of rules to be less harmful.

Google DeepMind

@GoogleDeepMind

2 years

Large language models can exhibit falsehoods, discriminatory language, and other unsafe behaviour. Introducing Sparrow: a dialogue agent that can search the internet and is trained to be more helpful, correct, and harmless using RL from human feedback: 1/

200

862

176

Maja Trebacz

@majatrebacz

3 years

We tuned a massive language model to support its answers with quotes from the web. I'm delighted to share this after putting much effort into it. Privileged to work on this project at @deepmind and collaborate with an amazing team @jacobmenick , @__nmca__ , @geoffreyirving et al.

Google DeepMind

@GoogleDeepMind

3 years

Introducing GopherCite, a fine-tuned version of Gopher that used human feedback to learn to back up claims with supporting evidence from the web. GopherCite can also answer questions about a given document or abstain when unsure. Learn more: 1/

137

616

Maja Trebacz

@majatrebacz

2 years

@MLinPL Really great event. It was a pleasure to attend and present. See you next year 🤙🇵🇱

Maja Trebacz

@majatrebacz

3 years

@jacobmenick Thank you, and right back at you! It was a great pleasure and learnt a ton from you :D