docs(local-llm): drop stale rep-penalty comparison for 80B escalation model

[JacobPEvans-personal]

Summary

• Removes the trailing "so it carries the same rep-penalty guardrail as the resident brain" clause from the 80B deep-escalation token-budget bullet in local-llm/mac-studio.mdx.
• The escalation model's own repetition_penalty 1.05 value is already stated earlier in the same sentence, so no information is lost.

Test plan

• jq empty docs.json
• mint broken-links — no broken links found
• ./scripts/validate-mermaid.sh — passes

https://claude.ai/code/session_0192RKGUtSAkf4f4WNrcQ5RE

[gemini-code-assist]

Code Review

This pull request updates the documentation in local-llm/mac-studio.mdx by removing a clause about the repetition penalty guardrail from the description of the 80B deep-escalation token budget. There are no review comments, and we have no additional feedback to provide.

[Copilot]

Pull request overview

This PR tightens the Mac Studio local-LLM documentation by removing a redundant clause in the 80B deep-escalation token budget bullet, keeping the stated guardrail facts intact while reducing repeated comparison phrasing.

Changes:

• Removes the trailing “same rep-penalty guardrail as the resident brain” clause from the 80B escalation token-budget bullet, since the escalation model’s repetition_penalty 1.05 is already stated in the same sentence.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.