experiment-plan

Name: experiment-plan
Availability: InStock
Author: duypham2801

Convert research proposals into claim-driven experimental roadmaps with ablation matrices and evaluation protocols.

Updated May 25, 2026

Search Tags:#experiment planning #research validation #ablation study #evaluation protocol #paper storyline #compute budget #ML baselines

Ready to run

npx skills add https://github.com/duypham2801/ThS_LLM --skill experiment-plan-duypham2801

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: experiment-plan
Source: https://github.com/duypham2801/ThS_LLM/tree/main/.claude/skills/experiment-plan
Command: npx skills add https://github.com/duypham2801/ThS_LLM --skill experiment-plan-duypham2801

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It converts a refined research idea into a concrete, reviewer-ready sequence of experiments that directly validate the paper’s key claims, including baselines, ablations, metrics, and an execution order.

Core Features & Use Cases

Claim-to-experiment mapping: turns a dominant thesis into specific claims with minimum convincing evidence and linked experiment blocks.
Paper storyline design: organizes experiments into main paper, appendix, and cut categories to keep the narrative compact and focused.
Execution-ready roadmap: produces a milestone run order (sanity → baseline → main method → decisive ablations → polish) with compute/data budget estimates and risk gates.
Structured outputs: writes a paper-oriented plan and a lightweight experiment tracker into refine-logs.

Quick Start

Use the experiment-plan skill to generate refine-logs/EXPERIMENT_PLAN.md and refine-logs/EXPERIMENT_TRACKER.md from your refined method and proposal notes, focusing on must-run claim-defending experiments.

Frequently Asked Questions about experiment-plan

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I design an experiment plan to defend my ML paper claims?▼

To design an experiment plan to defend your ML paper claims, you map your dominant thesis into specific claims, define the minimum convincing evidence, and link them to targeted experiment blocks including ablations and baselines.

What is claim-first experiment planning for machine learning research?▼

Claim-first experiment planning is a validation method that freezes your research claims before testing, ensuring every ablation matrix and evaluation protocol directly supports a specific contribution to keep your narrative compact and reviewer-ready.

How do I structure an ablation study and evaluation protocol under a compute budget?▼

You structure an ablation study under a compute budget by categorizing runs into must-run versus nice-to-have milestones, organizing experiments into main paper and appendix categories while applying risk gates to manage resource constraints.

Can I generate a run order for ML baselines and main method validation?▼

Yes, you can generate a run order for ML baselines and main method validation. The process creates a milestone execution sequence moving from sanity checks to baselines, then to the main method, decisive ablations, and final polish.

Does this experiment planning approach work for Diffusion and RL models?▼

Yes, this experiment planning approach works for Diffusion and RL models. It converts research proposals into claim-driven experimental roadmaps applicable across ML, VLM, LLM, Diffusion, and RL contributions for paper-ready validation.