Delentia SLM (v0.2.8-toon)
Small Language Model based on: Llama 3.1 8B
LoRA Rank (r)
32
LoRA Alpha
64
LoRA Dropout
0
Precision Mode
bf16 (T4/A100 optimized)
Performance Benchmarks
Recorded evaluation metrics validated through the Delentia OS pipeline:
JITNA Compliance94.2% (Target: > 94.0%)
FDIA Avg F-Score88.0% (Target: > 87.0%)
Hallucination Rate (lower is better)2.1% (Target: < 2.8%)
TOON Compliance92.0% (Target: > 90.0%)
Token Savings %46.5% (Target: > 15%)
Run Model Locally
If you have Ollama installed, run this model directly on your local system:
ollama run delentia-jitna-v0.2copy