fix(benchmark): correct postgres expected answers and remove inapplicable questions

[jafreck]

Summary

Fixes incorrect expected answers in the Postgres benchmark task definitions, removes questions that don't apply to the Postgres repo, and drops Q5.1 (semantic similarity) from both lore-self and postgres since it relies on embedding search rather than structural tools.

Changes

• Q1.6 (analyze.c dead-code question): Updated expected answer from 'None' to 'addNSItemForReturning'
• Q10.2 (top callers in utility.c): Updated with accurate caller counts from the SCIP index
• Removed Q2.1 (type hierarchy for TupleTableSlotOps): Not a meaningful type-hierarchy question for a C codebase
• Removed Q3.5 (external dependency mapping): Relies on build-system analysis outside Lore's current scope
• Removed Q5.1 (semantic similarity) from lore-self and postgres: Relies on embedding-based semantic search, which we want to exclude from the structural benchmark question set

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.49%. Comparing base (af76cab) to head (d1c9792).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #218   +/-   ##
=======================================
  Coverage   87.49%   87.49%           
=======================================
  Files          85       85           
  Lines        9475     9475           
  Branches     2932     2932           
=======================================
  Hits         8290     8290           
  Misses       1185     1185           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.