Search Results for author:
Found 10 papers, 6 papers with code
GPT-4 Technical Report
We report the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs.
Ranked #1 on
Long-Context Understanding
on Ada-LEval (BestAnswer)
Self-critiquing models for assisting human evaluators
On a topic-based summarization task, critiques written by our models help humans find flaws in summaries that they would have otherwise missed.
Training language models to follow instructions with human feedback
In this paper, we show an avenue for aligning language models with user intent on a wide range of tasks by fine-tuning with human feedback.
WebGPT: Browser-assisted question-answering with human feedback
This model's answers are preferred by humans 56% of the time to those of our human demonstrators, and 69% of the time to the highest-voted answer from Reddit.
Recursively Summarizing Books with Human Feedback
Our human labelers are able to supervise and evaluate the models quickly, despite not having read the entire books themselves.
Learning to summarize with human feedback
We collect a large, high-quality dataset of human comparisons between summaries, train a model to predict the human-preferred summary, and use that model as a reward function to fine-tune a summarization policy using reinforcement learning.
Learning to summarize from human feedback
We collect a large, high-quality dataset of human comparisons between summaries, train a model to predict the human-preferred summary, and use that model as a reward function to fine-tune a summarization policy using reinforcement learning.
Bayesian Inference of Regular Expressions from Human-Generated Example Strings
In programming by example, users "write" programs by generating a small number of input-output examples and asking the computer to synthesize consistent programs.
Pedagogical learning
samples are, in a sense, uninformative---they produce data without regard to how good this data is for learning.
Practical optimal experiment design with probabilistic programs
PPLs offer a clean separation between declaring problems and solving them, which means that the scientist can automate experiment design by simply declaring her model and experiment spaces in the PPL without having to worry about the details of calculating information gain.
