feat(bench): move most operation during dataset creation scripts

[lwaekfjlk]

Closes #

📑 Description

✅ Checks

• My pull request adheres to the code style of this project
• My code requires changes to the documentation
• I have updated the documentation as required
• All the tests have passed
• Branch name follows type/descript (e.g. feature/add-llm-agents)
• Ready for code review

ℹ Additional Information

[codecov]

Codecov Report

Attention: Patch coverage is 14.86486% with 126 lines in your changes missing coverage. Please review.

Project coverage is 78.02%. Comparing base (64502ec) to head (bc803a2).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
research_bench/run_eval.py	0.00%	53 Missing ⚠️
research_bench/utils.py	0.00%	35 Missing ⚠️
research_bench/mlbench.py	0.00%	16 Missing ⚠️
research_bench/crossbench.py	0.00%	12 Missing ⚠️
research_bench/proposal_writing.py	0.00%	6 Missing ⚠️
research_town/utils/paper_collector.py	77.77%	4 Missing ⚠️

@@            Coverage Diff             @@
##             main     #797      +/-   ##
==========================================
- Coverage   78.62%   78.02%   -0.61%     
==========================================
  Files          68       68              
  Lines        3102     3071      -31     
==========================================
- Hits         2439     2396      -43     
- Misses        663      675      +12     

Files with missing lines	Coverage Δ
research_town/data/data.py	100.00% <100.00%> (ø)
research_town/utils/error_handler.py	100.00% <100.00%> (ø)
research_town/utils/profile_collector.py	98.00% <100.00%> (+0.08%)	⬆️
tests/utils/test_paper_collector.py	100.00% <100.00%> (ø)
research_town/utils/paper_collector.py	67.50% <77.77%> (-3.73%)	⬇️
research_bench/proposal_writing.py	0.00% <0.00%> (ø)
research_bench/crossbench.py	0.00% <0.00%> (ø)
research_bench/mlbench.py	0.00% <0.00%> (ø)
research_bench/utils.py	0.00% <0.00%> (ø)
research_bench/run_eval.py	0.00% <0.00%> (ø)

... and 1 file with indirect coverage changes