Agents
RISED: Rubrics for Agentic Multi-Environment Selection and Self-Distillation
Training a single LLM agent jointly across diverse interactive environments has attracted increasing attention as a route to generalist agents. Existing curriculum and data-selection strategies often
Training a single LLM agent jointly across diverse interactive environments has attracted increasing attention as a route to generalist agents. Existing curriculum and data-selection strategies often allocate training at the environment level or prioritize local reward-based signals, without explicitly considering relationships between current rollouts across environments for prompt-group selection. Meanwhile, as environments are learned at different rates, all-failure and all-success rollout groups can coexist within a batch, leaving those data without group-relative reward signals. Both…
Related
- Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application
- pi-Play: Multi-Agent Self-Play via Privileged Self-Distillation without External Data
- AHEAD: Adaptive Hindsight with Environment-Augmented Distillation for Agentic RL
Source: Apple ML Research | 2026-10-06