Value-aligned language model fine-tuned for ethical reasoning through consequence analysis. Two-stage thinking training on the Swiss AI Apertus 8B Instruct base.

Approach

Karma Electric trains models on a structured ethical framework where the optimization target is suffering reduction rather than preference matching. Ethics emerges from understanding interdependence and consequences, not from learning surface-level preference patterns. For a full description of the framework see the Llama 3.1 8B release.

This Apertus variant uses the xIELU activation function (no gated MLP), enhanced multilingual pre-training, and Apertus-native <|inner_prefix|> / <|inner_suffix|> thinking tokens.

Current Version: v13

Two-stage training pipeline:

Stage 1 — Reasoning foundation (30k+ examples, 2 epochs)

Upstream extended thinking traces: Open-Orca, Dolphin, lordx64 Opus 4.7 reasoning distillation
Mixture-of-Thought (MoT) multi-domain reasoning

Stage 2 — KE ethics (~4,234 examples, 3 epochs)

Same Teapot-composed training data as v12
Consequence-based ethical reasoning with <think> traces converted to Apertus inner-monologue format

What's new in v13 (vs v12)

Two-stage training: reasoning foundation before ethics (v12 was ethics-only)
lordx64 Opus 4.7 reasoning traces in Stage 1 (3,500 high-quality extended thinking examples)
Richer upstream-thinking pool (30k+ vs none in v12)

Training details

QLoRA (4-bit NF4, bfloat16 compute, double-quant)
LoRA r=64, alpha=128, dropout 0.05, all attention and MLP projections (q, k, v, o, up, down)
Max context 4,096 tokens
Seed 42

Evaluation

Safety: 5/5 — refusals on weapons, phishing, Madhyamaka jailbreak, CSAM, social engineering
Sanity: 2/2 — coding and factual answers correct
Quality: 2/2 — substantive grief and career responses
Result: 9/10 on quick reward-probe (same as v12)

Safety

KE replaces refusal-template safety with consequence reasoning. The model holds boundaries by explaining real-world impact, not by citing policy.

Usage

Chat template

Apertus uses a native Jinja chat template with <|inner_prefix|> / <|inner_suffix|> for model-internal thinking. Use --jinja --chat-template-file with llama-server (or the equivalent Transformers apply_chat_template). The chat_template.jinja file is included in this repo.

llama.cpp

bash

# Conversation mode
llama-cli -m karma-electric-apertus-8b-v13-Q4_K_M.gguf -cnv \
    --jinja --chat-template-file chat_template.jinja

# Server mode
llama-server -m karma-electric-apertus-8b-v13-Q4_K_M.gguf \
    --port 8384 -c 4096 \
    --jinja --chat-template-file chat_template.jinja

Python (Transformers)

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "anicka/karma-electric-apertus-8b"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True)

messages = [
    {"role": "system", "content": open("system-prompt.txt").read().strip()},
    {"role": "user", "content": "How should I think about this ethical dilemma?"},
]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=800, do_sample=False)
print(tokenizer.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

System prompt

The recommended system prompt is in system-prompt.txt:

You are Karma Electric, an AI assistant grounded in ethical reasoning through consequence analysis and interdependence. You reduce suffering through honest, compassionate engagement — helping people see clearly while meeting them where they are. You maintain appropriate boundaries without moralizing or interrogating. Your goal is to reduce suffering, not to perform helpfulness.

Available Files

File	Description
model.safetensors	Merged model weights (bfloat16)
config.json, tokenizer.json, tokenizer_config.json	Standard Transformers files
chat_template.jinja	Apertus native chat template
karma-electric-apertus-8b-v13-Q4_K_M.gguf	Q4_K_M quantization for llama.cpp
system-prompt.txt	Recommended KE system prompt

Also Available

karma-electric-llama31-8b — Llama 3.1 8B v12, the primary release with full validation and activation-capping support.
karma-electric-qwen25-7b — Qwen 2.5 7B Instruct v12.
karma-electric-r1distill-llama-8b — DeepSeek R1-Distill-Llama-8B v12.

Project

Training scripts, datasets, and research documentation: github.com/anicka-net/karma-electric-project

Training composition tool: github.com/anicka-net/teapot

License

Apache 2.0 (Apertus base model license)