Self-Alignment with Instruction Backtranslation
We present a scalable method to build a high quality instruction followinglanguage model by automatically labelling human-written text with correspondinginstructions. Our approach, named instruction backtranslation, starts with alanguage model finetuned on a small amount of seed data, and a given…
Xian Li
Platypus: Quick, Cheap, and Powerful Refinement of LLMs
We present $\textbf{Platypus}$, a family of fine-tuned and merged LargeLanguage Models (LLMs) that achieves the strongest performance and currentlystands at first place in HuggingFace’s Open LLM Leaderboard as of the releasedate of this work. In this work we describe (1) our curated dataset$\tex…
Ariel N. LeeSolving Challenging Math Word Problems Using GPT-4 Code Interpreter with Code-based Self-Verification
Recent progress in large language models (LLMs) like GPT-4 and PaLM-2 hasbrought significant advancements in addressing math reasoning problems. Inparticular, OpenAI’s latest version of GPT-4, known as GPT-4 Code Interpreter,shows remarkable performance on challenging math datasets. In this paper…
Aojun ZhouTeach LLMs to Personalize -- An Approach inspired by Writing Education
Personalized text generation is an emerging research area that has attractedmuch attention in recent years. Most studies in this direction focus on aparticular domain by designing bespoke features or models. In this work, wepropose a general approach for personalized text generation using largel…
Cheng LiEfficient Guided Generation for Large Language Models
In this article we show how the problem of neural text generation can beconstructively reformulated in terms of transitions between the states of afinite-state machine. This framework leads to an efficient approach to guidingtext generation with regular expressions and context-free grammars by al…
Brandon T. Willard