[Sampling Replay 8/8] Explain sampling-support replay

[nanjiangwill]

Stack position

Part 8/8 of the bounded sampling-support replay series. Merge after #3352. #3321 is merged in main.

Layer-only diff: nanjiangwill/miles@test/sampling-replay-e2e...docs/sampling-support-replay

Review umbrella: #3354.

Summary

• Explain why bounded rollout sampling and full-vocabulary actor normalization differ.
• Define the fixed-support replay objective and configuration.
• Document the Miles and SGLang ownership boundary.
• Document current temperature, router, prefill, penalty, KL, and distillation constraints, including that the v0.5.20 router limitation is request-side typed serialization rather than response stripping.
• Add the feature to Advanced navigation and both CLI sampling references.
• Explain how to interpret rollout/trainer log-probability mismatch metrics after the first update.

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[nanjiangwill]

Closing this layer-specific review PR in favor of the single integrated sampling-replay PR #3354. The implementation and validation remain available there.