📜 💡**New paper & dataset** Our paper [to be presented as a spotlight at NeurIPS’25] presents DIVE, a novel dataset for aligning GenAI models to pluralistic viewpoints. DIVE and accompanying experiments advocate for a shift in both: alignment strategies and safety evaluations.
🤔🤖Most AI systems assume there’s just one right answer—but many tasks have reasonable disagreement. How can we better model human variation? 🌍✨
We propose modeling at the individual-level using open-ended, textual value profiles! 🗣️📝
https://t.co/PUr3olk5u6
(1/?)
This is huge: Llama-v2 is open source, with a license that authorizes commercial use!
This is going to change the landscape of the LLM market.
Llama-v2 is available on Microsoft Azure and will be available on AWS, Hugging Face and other providers
Pretrained and fine-tuned models are available with 7B, 13B and 70B parameters.
Llama-2 website: https://t.co/PKrrXgHdem
Llama-2 paper: https://t.co/aINNrXNhMb
A number of personalities from industry and academia have endorsed our open source approach: https://t.co/N7HwgW9Suh
This is huge: Llama-v2 is open source, with a license that authorizes commercial use!
This is going to change the landscape of the LLM market.
Llama-v2 is available on Microsoft Azure and will be available on AWS, Hugging Face and other providers
Pretrained and fine-tuned models are available with 7B, 13B and 70B parameters.
Llama-2 website: https://t.co/PKrrXgHdem
Llama-2 paper: https://t.co/aINNrXNhMb
A number of personalities from industry and academia have endorsed our open source approach: https://t.co/N7HwgW9Suh
Come and join us at @AmsterdamNLP! We have a PhD position in #NLProc with a focus on modelling the spread of misinformation, in collaboration with Meta AI. Looking for students with an NLP/ML background and an interest in language and society. https://t.co/6d42RiWb2b
Attending my first #NeurIPS2021🤩
I'll be presenting our work
"Towards a robust experimental framework & benchmark for lifelong language learning"
my #AI Master thesis with
@HYannakoudakis@Nithin_Holla@ImPushMish@KatiaShutova
📢1430-1800 GMT Dec 7
🔗 https://t.co/x5WXAbd2dx
Keeping @Facebook users safe from policy violating content like hate speech, misinformation or child exploitation, is a top priority for @facebookai. This paper surveys the AI techniques we have developed and used to preserve integrity on our platform https://t.co/rHniYXS5rc
Modelling online communities using graph neural networks for abusive language detection. Our #naacl2019 paper is now on arXiv! https://t.co/gxWseYY1Xs #NLProc with @pushkarmishra96 @marcodeltredici@KatiaShutova
@Tuhin66978276 @HYannakoudakis@marcodeltredici@KatiaShutova The racist tweets in the dataset indeed come from very few authors (only 5). That said, the racist authors (including the one you talk about) have contributed many sexist and clean tweets too. This camouflages their racist behavior to some extent. We improve greatly on sexism too
“Abusive language detection using graph convolutional networks” with @pushkarmishra96 @marcodeltredici Katia Shutova accepted at #naacl2019! See you in Minneapolis!
@danieljkelley The dataset we used to bootstrap the AI contains abusive examples of racism against muslims. Sexist examples are limited to a specific scenario. The dataset itself is quite small sadly. Hence we seek more data :) I have DMed you the link to dataset.