feat(hints, summary, docs): Refactor O3 parameters to generic names and restructure GAIA validation docs

[BinWang28]

Summary

• Renamed O3-specific parameters to generic model-agnostic names across all configs and code
• Extracted common GAIA validation content into gaia_validation_prerequisites.md
• Renamed config files for consistency: agent_gaia-validation_mirothinker.yaml, agent_gaia-validation_claude37sonnet.yaml
• Added multiple runs section for MiroThinker with performance variance explanation
• Moved API keys configuration to model-specific documentation

Changes

• Parameters: o3_hint → hint_generation, o3_final_answer → final_answer_extraction, etc.
• Code: Updated function names and references in orchestrator.py and summary_utils.py
• Configs: Updated all YAML files with new parameter names
• Docs: Restructured GAIA validation documentation with model-specific files
• Scripts: Updated script references to use new config file names

[Copilot]

Pull Request Overview

This pull request refactors O3-specific parameter names to generic model-agnostic names and restructures GAIA validation documentation. The changes make the codebase more flexible and maintainable by removing model-specific naming conventions.

Key changes include:

• Renamed parameters: o3_hint → hint_generation, o3_final_answer → final_answer_extraction
• Updated function names and references throughout the codebase
• Extracted common GAIA validation content into separate documentation files
• Added support for configurable base URLs for hint generation and final answer extraction

Reviewed Changes

Copilot reviewed 35 out of 36 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
utils/util_llm_simple_voting.py	Updated string replacement to use generic "LLM" instead of "O3"
utils/util_llm_parallel_thinking.py	Updated string replacement and improved function parameter formatting
src/utils/summary_utils.py	Renamed functions to generic names and added base_url parameters
src/core/orchestrator.py	Updated function calls and variable names to use generic terminology
config/*.yaml	Updated all configuration files to use new parameter names
docs/mkdocs/docs/*.md	Restructured GAIA validation docs and updated references
scripts/*.sh	Updated script to use new config file names

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Remove the unnecessary return statement at the end of the generator function. Generator functions don't need explicit empty return statements.

Suggested change

return

[Copilot]

Remove the commented-out debug code. This appears to be leftover debugging code that should be cleaned up.

Suggested change

# response = await client.chat.completions.create(messages = [{"role": "user", "content": content}], model="dummy")

[Copilot]

Remove the commented-out parameter. If this parameter is no longer needed, the commented line should be removed to keep the code clean.

Suggested change

# reasoning_effort="medium",

[Copilot]

Pull Request Overview

Copilot reviewed 35 out of 36 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

This commented-out line appears to be debug/test code that should be removed from production code.

Suggested change

# response = await client.chat.completions.create(messages = [{"role": "user", "content": content}], model="dummy")

[Copilot]

This commented-out line should be removed as it's redundant with the active line below it.

Suggested change

# client = AsyncOpenAI(api_key=api_key, timeout=600)

[Copilot]

This commented-out parameter should either be removed or properly documented if it's intentionally disabled for generic LLM compatibility.

Suggested change

	messages=[{"role": "user", "content": f"[{message_id}] {full_prompt}"}],
	messages=[{"role": "user", "content": f"[{message_id}] {full_prompt}"}],
	# The 'reasoning_effort' parameter is intentionally disabled for compatibility with generic LLM APIs.
	# Uncomment and set as needed if using an LLM that supports this parameter, e.g., OpenAI's proprietary models.