Excited to receive an ACL outstanding paper award, with @max_nlp @latticecut @riedelcastro @ucl_nlp ! TL;DR If prompting is not working, change the order, the performance may jump from random-guess to SOTA. How to find fantastically ordered prompts? Here➡️ https://t.co/R6NktEebwJ

🚨 New Paper 🚨 Can LLMs perform latent multi-hop reasoning without exploiting shortcuts? We find the answer is yes – they can recall and compose facts not seen together in training or guessing the answer, but success greatly depends on the type of the bridge entity (80%+ for

🎉🎉🏆

They say a picture is worth a thousand words... but work led by @ralph_tang finds words worth a thousand pictures! https://t.co/L8wU715wRd

🍷 FineWeb technical report is out and so is 📚 FineWeb-Edu, a 1.3 trillion tokens dataset that outperforms all other open web datasets, with remarkable improvements on educational benchmarks such as MMLU, ARC, and OpenBookQA. Technical report: https://t.co/lfOZYYJKxq Dataset:

@ralph_tang @riedelcastro (3/3) 📈 Compared to the standard templates widely used in LLM benchmarks, our random prompt baseline yielded an average relative improvement of over 10%! Take a look at our work https://t.co/MHac6nMg2J, unlocking the potential of language model optimisation space!

@ralph_tang @riedelcastro (2/3) we're outperforming human experts! This establishes a robust baseline for prompt optimisation. Even for basic text classification tasks, Large Language Models (such as LLaMA, Mistral, ChatGPT) can experience significant improvement using #python random prompt optimisation

(1/3) Excited to introduce our #NAACL2024 paper "Strings from the Library of Babel: Random Sampling as a Strong Baseline for Prompt Optimisation". We've unlocked the power of #python random library for prompt optimisation✨With a simple trick - random.sample(vocabulary) as prompt

We're excited to announce AIDE has become the first human-level AI agent for data science! AIDE outperforms half of human data scientists on a wide range of Kaggle competitions, surpassing conventional AutoML, LangChain agents, and ChatGPT with human assistance. 🏆

Congrats on the launch! LFG @YuxiangJWu @zhengyaojiang

Often prompt engineering focuses on the *content* of the prompt, but in reality *formatting* of the prompt can have an equal or larger effect, especially for less powerful models. This is a great deep dive into this phenomenon by @melaniesclar et al.

LLMs are now trained >1000x as much language data as a child, so what happens when you train a "BabyLM" on just 100M words? The proceedings of the BabyLM Challenge are now out along with our summary of key findings from 31 submissions: https://t.co/zli0jzA1XP Some highlights 🧵

David Attenborough is now narrating my life Here's a GPT-4-vision + @elevenlabsio python script so you can star in your own Planet Earth:

🚨💫We are delighted to have @shishirpatil_ at our @uclcs NLP Meetup *Monday 1st Nov 6:30pm GMT* The event will be *hybrid* Due to room capacity, there are *two links* to sign up depending on whether you attend in person or online Details in: https://t.co/pcnWPmDt8F

Introducing ChatArena 🏟 - a Python library of multi-agent language game environments that facilitates communication and collaboration between multiple large language models (LLMs)! 🌐🤖 Check out our GitHub repo: https://t.co/xisQrxeQr7 #ChatArena #NLP #AI #LLM 1/8 🧵

We would like to wish you all a restful winter break, and best wishes for the new year 🙂🎄

Happy to share MARGE, our new work on rethinking pre-training: given a document, we first retrieve related documents, and then paraphrase these to reconstruct the original. MARGE works well for generation and classification in many languages, sometimes without supervision. (1/6)

We will present a tutorial on Machine Learning for Drug Development at #IJCAI2020! Materials to follow on our website: https://t.co/w0JZJhef8B @IJCAIconf #drugs #networks #AI

#ICLR2020 paper on Differentiable Reasoning over a Virtual Knowledge Base: Efficient, end-to-end differentiable framework for doing complex multi-hop QA over a large text corpus. https://t.co/P60JiWv3EC w/t Dhingra, Zaheer, Balachandran, @gneubig , @professorwcohen

Transformers as Soft Reasoners over Language "we explore whether transformers can similarly learn to reason (or emulate reasoning), but using rules expressed in language, thus bypassing a formal representation." https://t.co/5uib4bRPud Datasets and demo: https://t.co/wKxgdCpHiE