 Claude Code
|
ba4a2f4262
|
Switch all Ollama models to gemma3n:e4b on node-197 GPU
Bake-off results: gemma3n:e4b (80.6% cmd match, 100% safety, 5.9s)
outperforms qwen3-coder:30b on all metrics. Updated paper, shrink,
and langgraph gateway configs. Frees steel141 for LoRA training.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
2026-03-18 10:29:54 -04:00 |
|
 Seth
|
8ee8be9cc0
|
Initial commit — Minecraft AI God plugin
- mc_aigod.py: main watcher script (log tail, RCON, two-call LLM)
- Two-call LLM split: qwen3-coder:30b for commands, gemma3:12b for messages
- Divine intervention timer (Poisson process, configurable avg/day)
- Prayer memory (persistent, last 10 exchanges)
- Rolling server log context (last 20 min events)
- Live player context (inventory with rarity, health, food, pos, XP)
- /pray and bible chat detection (no slash — vanilla 1.21 compatible)
- Login notice, bible help system
- debug_commands toggle (in-game command display via tellraw)
- Auto-fix for transposed give command syntax
- JSON repair fallback for truncated LLM responses
- Sentence-aware message chunking for long responses
- mc-aigod.service systemd unit
- mc_aigod_shrink.json example config
- README.md full implementation guide
- Minecraft_Ai_God.md full design document
|
2026-03-15 19:02:16 -04:00 |
|