Rohan Taori
Pretraining researcher at Anthropic; co-author of Stanford Alpaca, AlpacaFarm and AlpacaEval
Rohan Taori is a researcher on the pretraining team at Anthropic. Taori received a PhD in computer science at Stanford, advised by Tatsu Hashimoto, and is one of the five graduate-student authors who contributed equally to Stanford Alpaca, an instruction-following model fine-tuned from a 7B LLaMA model. Taori also co-authored AlpacaFarm, a simulation framework for methods that learn from human feedback, and AlpacaEval, an automatic evaluator for instruction-following language models. Earlier work listed on Taori's site includes a study of how image classifiers hold up under natural distribution shifts.1345
Work
Stanford Alpaca (external site: github.com)Open-source code
Listed in the repository's Authors section among the graduate students who contributed equally, all advised by Tatsunori B. Hashimoto. The code is Apache-2.0; the README says the dataset and weight diff are CC BY NC 4.0 and intended for research use only.3
AlpacaFarm (external site: arxiv.org)Published research
Co-author of "AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedback" (2023); Taori's site marks Taori as one of four equal-contribution authors.41
AlpacaEval (external site: github.com)Open-source code
Listed on Taori's site as an equal-contribution co-author of AlpacaEval, an automatic evaluator for instruction-following language models. The tatsu-lab/alpaca_eval repository is Apache-2.0.15
Other documented roles
Roles the sources document but that are not confirmed as current.
- PhD student in Computer Science, advised by Tatsu Hashimoto, Stanford University13
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.