MLC: Open Collab is a 100% open community for independent researchers. We held our tenth (!!!) Research Jam on August 24, 2022, where four presenters signed up to share updates on their ongoing research. Access slides and read more about it at mlcollective.org/research-jam...
0:00 Event starts
01:01 Examining gradients of generalization in RL agents
13:54 A step towards understanding the accuracy distribution
24:57 "The fire alarm just went off"
25:47 `gymnax`: RL environments in JAX
39:32 Prompt forgetting