code/Taotern_LLM_Experiments/docs/showcase/README.md

# Showcase Documentation Guide

This folder contains the current showcase-facing summary of the Taotern GSM/SSM LLM project.

Recommended reading order:

1. GSM_RnD_Showcase_Report.md

   Start here. This is the full technical story from CHASM context to GSM, Euler Gamma SSM, bilinear/S4-style Gamma with convolution, DPLR, TaoNet SSM, hybrid TaoNet, and the 200M model candidates. It includes what worked, what failed, current benchmark findings, and open questions.

2. GSM_Scaling_Law_Notes.md

   Read this after the full report if you want the short paper-style lessons. It focuses on empirical scaling rules such as attention-to-SSM ratio, projection dimension, local shift, split lanes, token budget, and chatbot-readiness steps.

3. GSM_RnD_Showcase_Report.docx

   This is the Word version of the full report. Use it when preparing a paper draft, presentation material, or a document for non-code review.

Suggested usage:

• For technical onboarding: read the full Markdown report first.
• For paper writing: use the .docx plus the scaling-law notes.
• For quick project positioning: read the executive summary and scaling-law notes.

The Markdown files are easier to maintain in Git. The .docx should be regenerated from the full report when major content changes are made.