lattice-ai/dolphin-mistral-nemo-12b

Lattice Dolphin Mistral Nemo 12B

Privacy Tier: wrapped | Parameters: 12.2B | Context: 131,072 tokens | VRAM: ~24GB

Dolphin on Mistral Nemo 12B — the sweet spot for uncensored reasoning. 12B parameters hit the quality threshold where fine-tuned models start to genuinely surprise you. Apache 2.0 for full commercial freedom.

Privacy Guarantees

Feature	Status
Sandboxed training (no network egress)	Yes
PII output guardrails	Yes
Encrypted training logs	Yes
Zero telemetry	Yes
DP-SGD training support	Yes
Privacy certificate on export	Yes

Quick Start

pip install ltce
ltce pull lattice-ai/dolphin-mistral-nemo-12b
ltce train ./your-data --model dolphin-mistral-nemo-12b --epsilon 4.8 --method qlora
ltce verify ./output/adapter

from ltce import Lattice

lt = Lattice()
vault = lt.encrypt("./sensitive-data/", password="...")
result = lt.train(
    model="dolphin-mistral-nemo-12b",
    data=vault,
    epsilon=4.8,
    method="qlora",
)
lt.verify(result)

What is Lattice?

Lattice is a privacy-first model training platform. The value isn't running locally (anyone can do that). The value is:

• DP-SGD training -- individual training examples can't be extracted from weights
• Signed certificates -- BLAKE3 hash + ed25519 signature proves the privacy guarantee
• Safe sharing -- publish your adapter knowing the training data is mathematically protected

Capabilities

general, instruct, reasoning, uncensored

Base Model

cognitivecomputations/dolphin-2.9.4-mistral-nemo-12b

License

Apache 2.0

---

Built with Lattice -- Train private. Prove it. Share safely.