Collections for the paper "Language Models Can Learn from Verbal Feedback Without Scalar Rewards" (https://arxiv.org/pdf/2509.22638)