A General Language Assistant as a Laboratory for Alignment

By Amanda Askell, Yuntao Bai, Anna Chen, Dawn Drain, Deep Ganguli, T. Henighan, Andy Jones, Nicholas Joseph, Benjamin Mann, Nova Dassarma and 12 more. 2021

1,175 citations recorded by the source corpus.

Given the broad capabilities of large language models, it should be possible to work towards a general-purpose, text-based assistant that is aligned with human values, meaning that it is helpful, honest, and harmless. As an initial foray in this direction we study simple baseline techniques and evaluations, such as prompting. We find that the benefits from modest interventions increase with model size, generalize to a variety of alignment evaluations, and do not compromise the performance of large models. Next we investigate scaling trends for several training objectives relevant to alignment, comparing imitation learning, binary discrimination, and ranked preference modeling. We find that ranked preference modeling performs much better than imitation learning, and often scales more favorably with model size. In contrast, binary discrimination typically performs and scales very similarly to imitation learning. Finally we study a `preference model pre-training' stage of training, with the goal of improving sample efficiency when finetuning on human preferences.

Subfields: Artificial Intelligence, Computer Networks and Communications, Law, Large Language Models, Alignment & RLHF

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

  1. Mohit Bansal
  2. Jianfeng Gao
  3. Luke S. Zettlemoyer
  4. Percy Liang
  5. Yejin Choi
  6. Yoshua Bengio
  7. Quoc V. Le
  8. Kyunghyun Cho

In Computer Networks and Communications

  1. Pieter Abbeel
  2. Matthew Botvinick
  3. Sergey Levine
  4. Jacob Steinhardt
  5. Dan Hendrycks
  6. Stanislav Fort
  7. Percy Liang
  8. Fan Yang

In Law

  1. Bertie Vidgen
  2. Jamin Shin
  3. William S. Isaac
  4. Iason Gabriel
  5. Harrison Edwards
  6. Yong Jae Lee
  7. Percy Liang
  8. Maosong Sun

In Large Language Models

  1. Luke S. Zettlemoyer
  2. Hinrich Schütze
  3. Yejin Choi
  4. Hannaneh Hajishirzi
  5. Danqi Chen
  6. Jianfeng Gao
  7. Percy Liang
  8. Samuel R. Bowman

In Alignment & RLHF

  1. Geoffrey Irving
  2. Ethan Perez
  3. Danqi Chen
  4. Jan Leike
  5. Jacob Hilton
  6. Wei Xiong
  7. Tomasz Korbak
  8. Jacob Steinhardt

Citations in the corpus