Seth/Mortdecai
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Minecraft knowledge corpus, recipe trees, GitHub scraper, 644 examples

Browse Source 
Knowledge corpus (knowledge/mc-data/):
- 1505 items, 886 crafting recipes, 1166 blocks from minecraft-data 1.21.11
- Recipe dependency tree builder (knowledge/build_recipe_tree.py)
- Crafting chain training: "give me everything to make X from scratch"
- Smelting recipes, version awareness examples

Training data (644 examples total):
- 107 command syntax reference examples (every command + common errors)
- 176 recipe/crafting chain examples (63 crafting, 103 material-giving, 11 smelting)
- 344 Claude-distilled examples (222 sudo + 122 god via Haiku)
- Live bot audit data ingested (128 examples from dev server)

Swarm bots:
- Swimming/water escape logic
- Door opening
- Context-aware prayers (inventory, health, time, depth)
- Prefix enforcement on all Gemini/Dolphin prompts

GitHub log scraper (data/scrape_server_logs.py):
- Searches GitHub for Minecraft server logs with commands
- Strict 1.20.5+ version filter
- Extracts command pairs, converts to training format

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
Seth Freiberg
2026-03-18 20:33:09 -04:00
parent 65ee146043
commit 0473eb0b50
14 changed files with 103586 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
PLAN.md
+3
View File
@@ -367,6 +367,9 @@ These are ideas to explore after the core system is working. Prioritize based on
| 2026-03-18 | First LoRA training run (233 examples, 3 epochs) | Loss 1.5→0.10. Model is bad — hallucinating Chinese, leaking system prompt. Expected at this data scale. Deployed to dev server for live data collection. |
| 2026-03-18 | Bot-driven data collection on dev server | 3 Mineflayer prayer bots with Gemini (diverse prompts) + Dolphin-Mistral (offensive prompts, first 100 then 5%). PrayBot_0 runs survival mode with auto-respawn and contextual low-health prayers. |
| 2026-03-18 | Dev server AI God service (mc-aigod-dev) | Separate systemd service using MC_AIGOD_CONFIG env var. Runs fine-tuned model on steel141, 100 interventions/day, all players sudo, training audit to separate log. |
| 2026-03-18 | Minecraft knowledge corpus baked into training | 1505 items, 886 recipes, 1166 blocks from minecraft-data 1.21.11. Recipe dependency trees, smelting knowledge, crafting chain examples. 107 command ref + 176 recipe examples. |
| 2026-03-18 | Claude distillation: God Soul + Haiku | God Soul document adapted from Claude's soul framework. Haiku distills 344 training examples ($0.65). Dev server switched to Haiku API ($5 budget) for high-quality live data. |
| 2026-03-18 | Version-aware training | Model trained to know it targets 1.21.x, understands 1.20.5 syntax changes, knows recipes evolve with updates. |
---