https://sumust.github.io/2026-07-07 https://sumust.github.io/blog.html2026-07-07 https://sumust.github.io/publications.html2026-07-07 https://sumust.github.io/blog/grpo/2026-07-07 https://sumust.github.io/blog/two-kls/2026-07-07 https://sumust.github.io/blog/trust-region-subproblem/2026-07-07 https://sumust.github.io/blog/trust_regions/2026-07-07 https://sumust.github.io/blog/rosenbrock-explorer/2026-07-07 https://sumust.github.io/blog/content/rosenbrock-explorer.html2026-07-07 https://sumust.github.io/blog/content/subproblem-explorer.html2026-07-07 https://sumust.github.io/blog/content/bandit-explorer.html2026-07-07 https://sumust.github.io/blog/content/kl-penalty-explorer.html2026-07-07 https://sumust.github.io/blog/content/baseline-explorer.html2026-07-07