I was also happy to see that this paper, and an earlier one by Hase et al. also on counterfactual simulatability (but more focused on training) used tinker for their fine-tuning experiments
Original titleI was also happy to see that this paper, and an earlier one by Hase et al. also on counterfactual simulatability (but more focused on tra...
Read the original x.comSource: John Schulman · x.comPublished · added here