Tim Rocktaschel
28 Papers
5 Citations
Tim Rocktaschel is an academic researcher. The author has contributed to research in topics: Computer science & Medicine. The author has an hindex of 5, co-authored 13 publications.
Chat about Author
Papers
Genie: Generative Interactive Environments
Jake Bruce,Michael Dennis,Ashley Edwards,Jack Parker-Holder,Yuge Shi,Edward Hughes,Matthew Lai,Aditi Mavalankar,Richie Steigerwald,Chris Apps,Yusuf Aytar,Sarah Bechtle,Feryal Behbahani,Stephanie Chan,Nicolas Heess,Lucy Gonzalez,Simon Osindero,Sherjil Ozair,Scott Reed,Jingwei Zhang,Konrad Żołna,Jeff Clune,Nando de Freitas,Satinder Singh,Tim Rocktaschel +24 more
TL;DR: Genie is introduced, the first generative interactive environment trained in an unsupervised manner from unlabelled Internet videos, which enables users to act in the generated environments on a frame-by-frame basis despite training without any ground-truth action labels or other domain-specific requirements typically found in the world model literature.
39
Debating with More Persuasive LLMs Leads to More Truthful Answers
Akbir Khan,John Hughes,Dan Valentine,Laura Ruis,Kshitij Sachan,Ansh Radhakrishnan,Edward Grefenstette,Samuel R. Bowman,Tim Rocktaschel,Ethan Perez +9 more
TL;DR: It is found that debate consistently helps both non-expert models and humans answer questions, and optimising expert debaters for persuasiveness in an unsupervised manner improves non-expert ability to identify the truth in debates.
29
Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks
Samyak Jain,Robert Kirk,Ekdeep Singh Lubana,Robert P. Dick,Hidenori Tanaka,Edward Grefenstette,Tim Rocktaschel,D. Krueger +7 more
TL;DR: It is shown that practitioners can unintentionally remove a model's safety wrapper merely by fine-tuning it on a, e.g., superficially unrelated, downstream task.
28
Large language models are not zero-shot communicators
TL;DR: A simple task is designed and widely used state-of-the-art models are evaluated, finding that, despite only evaluating on utterances that require a binary inference (yes or no), most perform close to random.
Efficient Planning in a Compact Latent Action Space
Zheng Qiang Jiang,Tianjun Zhang,Michael Janner,Yueying Li,Tim Rocktaschel,Edward Grefenstette,Yuandong Tian +6 more
- 22 Aug 2022
TL;DR: The Trajectory Autoencoding Planner is proposed, a planning-based sequence modelling RL method that scales to high state-action dimensionalities and surpasses existing model-based methods, including TT, with a large margin and also beats strong model-free actor-critic baselines.
25