Skip to content

Missing details for SFT data pipeline reproduction #36

Description

@haozhihuiYmh150

To reproduce the SFT data pipeline (Section 3-4), I'm missing three things:


  1. IssueTasks data source (Table 2, rank 4) — no generate.py in data/. How was it generated?
    
  2. Tezos 997→21K instruction rewriting prompt (Section 4) — Found HARDEN_PROMPT in data/synthetic_harden/utils.py (gpt-4o-mini, temp=0.7), which looks like the right direction. 🙏 One clarification: this prompt focuses on "hardening" (adding constraints/complexity) — was the 997→21K expansion done by running harden_instructions() multiple times with varied seeds, or was there a separate diverse rephrasing step with a different prompt? If the former, then HARDEN_PROMPT × multiple invocations at temp=0.7 already answers it.
    
  3. Response-length upsampling formula (Section 4) — "proportionally to score" is underspecified. What's the exact formula mapping response-length to rollout count? Any cap/min/max? Total rollout budget?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions