Published December 8, 2023 | Version v1
Model Open

Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study (models and test data)

Authors/Creators

  • 1. ROR icon University of Bath

Description

Trained models (logs.zip) and test data (experiments.zip) for NeurIPS 2023 ATTRIB workshop paper "Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study".

Paper: https://arxiv.org/abs/2312.03762

Code: https://github.com/KarolisRam/colour-shape-goal-misgeneralization

Files

experiments.zip

Files (7.5 GB)

Name Size
md5:07dd1f47a2ec30c53b9d08051d2ae0c3
81.2 MB Preview Download
md5:74a9a4b6bfd44ca8f65b4a74ba1c1dfe
7.4 GB Preview Download
md5:c227a82e0fd5972abe7b20eb3f9c1c3d
247 Bytes Preview Download

Additional details

Dates

Submitted
2023-12-08