GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
saidutta69
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
saidutta69
Model Tree
utter-project/EuroLLM-9B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Your GPU | Recommended quant | Weights |
|---|---|---|
| RTX 3090 / 4090 / 5090 (24 GB) | Q8_0 | 9.73 GB |
| RTX 4080 / 5080 (16 GB) | Q6_K | 7.51 GB |
| A6000 / A100 40GB (40 GB) | Q8_0 | 9.73 GB |
| CPU-only (32 GB RAM) | Q4_K_M | 5.58 GB |
Weights only, at this model's 9.15B native size. At 9B this needs a real GPU - it will run on CPU but not quickly. Note the short context: KV cache is small, so this model is compute-bound rather than memory-bound on long generations.
Trial 100 of a 200-trial Heretic run (seed 2494858073).
| Parameter | Value |
|---|---|
| direction_index | 23.66 |
| attn.o_proj.max_weight | 0.88 |
| attn.o_proj.max_weight_position | 27.67 |
| attn.o_proj.min_weight | 0.83 |
| attn.o_proj.min_weight_distance | 18.96 |
| mlp.down_proj.max_weight | 1.16 |
| mlp.down_proj.max_weight_position | 35.07 |
| mlp.down_proj.min_weight | 1.12 |
| Metric | This model | Original model (utter-project/EuroLLM-9B-Instruct) |
|---|---|---|
| KL divergence | 0.0383 | 0 (by definition) |
| Refusals | 4/100 | 100/100 |
The base model refused every single prompt in the harmful evaluation set - 100/100, the most complete refusal tuning of any model in this collection - and that is reduced to 4/100. The KL divergence of 0.0383 is a moderate edit: it sits between the very tight runs (Qwen3-4B-Thinking-heretic at 0.0011) and the deliberately loose one (granite-4.1-3b-heretic at 0.0852), so expect some drift in phrasing alongside the refusals disappearing.
Abliteration does not add capability. The base model's evaluation tables are published as images, so only the summary claims are quotable here - no invented numbers.
Fine-tuning a "helpful" persona on top of RLHF'd refusals fights the base model's training and tends to degrade coherence. Abliteration instead finds and edits the specific weight directions responsible for refusal, leaving the rest of the network (and its capabilities) untouched. See the Heretic repo and the original abliteration writeup for the mechanism.
Made with ❤️ by RACER IS OP — follow for more uncensored models
| File | Size |
|---|---|
model-00001-of-00004.safetensors | 5.00 GB |
model-00002-of-00004.safetensors | 4.92 GB |
model-00003-of-00004.safetensors | 4.92 GB |
model-00004-of-00004.safetensors | 3.47 GB |
BF16, 9,152,319,488 parameters (9.15B), with untied input/output embeddings
(~0.52B each). The reproduce/ directory carries the full Heretic recipe - config.toml,
requirements.txt, the Optuna study journal, and SHA-256 sums - so this exact model can be
regenerated bit-for-bit. Reproduce it with heretic --reproduce reproduce/reproduce.json.
Default set (F16 + 4 quants) produced with llama.cpp.
| File | Format | Size |
|---|---|---|
EuroLLM-9B-Instruct-heretic-F16.gguf | GGUF F16 | 18.31 GB |
EuroLLM-9B-Instruct-heretic-Q4_K_M.gguf | GGUF Q4_K_M | 5.58 GB |
EuroLLM-9B-Instruct-heretic-Q5_K_M.gguf | GGUF Q5_K_M | 6.52 GB |
EuroLLM-9B-Instruct-heretic-Q6_K.gguf | GGUF Q6_K | 7.51 GB |
Llama architecture (llama, 381 tensors, 42 layers) - loads natively in
llama.cpp / Ollama / LM Studio / Jan, and is a standard dense model so MLX and vLLM work too.
Run llama serve -hf saidutta69/EuroLLM-9B-Instruct-heretic to pull the default quant.
bash
# llama.cpp -- override the short native context if you have the RAMllama serve -hf saidutta69/EuroLLM-9B-Instruct-heretic -c 8192python
# transformersfrom transformers import AutoModelForCausalLM, AutoTokenizer model_name = "saidutta69/EuroLLM-9B-Instruct-heretic"model = AutoModelForCausalLM.from_pretrained(model_name, dtype="auto", device_map="auto")tokenizer = AutoTokenizer.from_pretrained(model_name) messages = [ {"role": "system", "content": "You are EuroLLM --- an AI assistant specialized in European languages."}, {"role": "user", "content": "Translate to Polish: 'The library closes at six.' Also answer in English: what is the capital of Portugal?"},]inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt").to(model.device)out = model.generate(**inputs, max_new_tokens=1024)print(tokenizer.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))Abliteration does not touch the tokenizer, so the base model's multilingual setup is intact. Note the system prompt above is the one the base model card recommends - keep a language hint in it if you want reliable language selection.
| Architecture | LlamaForCausalLM (decoder-only dense transformer) |
| Parameters | 9.15B (0.52B embeddings + 0.52B LM head + 8.10B non-embedding) |
| Layers / heads | 42 layers, 32 attention heads, 8 KV heads (GQA), head_dim 128 |
| Hidden / intermediate | 4096 / 12288 (SwiGLU, pre-layer RMSNorm) |
| Position embedding | RoPE, theta = 10,000 |
| Context length | 4,096 native - no rope_scaling in the config |
| Vocab | 128,000 |
Refusal suppression is deliberate and works as intended: this model will comply with requests the base model would refuse, including some it shouldn't. There is no safety filtering layered on top. You are responsible for how you deploy it - don't put this behind an unmoderated public-facing endpoint serving third parties. It inherits EuroLLM's factual limitations and biases across all 35 languages; abliteration removes refusal directions, it doesn't add capability or judgment.
Because this model is strong at translation into and out of many languages, treat it accordingly: unfiltered translation of untrusted content is a good way to launder text past another system's filters.
Inherits the apache-2.0 license from the base
model, which is tagged Apache-2.0 on the Hub but ships no LICENSE file, so the canonical license
text is linked rather than a repo blob.
| mlp.down_proj.min_weight_distance | 23.46 |
EuroLLM-9B-Instruct-heretic-Q8_0.gguf | GGUF Q8_0 | 9.73 GB |
| Precision | bfloat16, untied embeddings |
| Languages | Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Irish, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Slovak, Slovenian, Spanish, Swedish, Arabic, Catalan, Chinese, Galician, Hindi, Japanese, Korean, Norwegian, Russian, Turkish, Ukrainian |
| Developers | Unbabel, Instituto Superior Tecnico, Instituto de Telecomunicacoes, University of Edinburgh, Aveni, University of Paris-Saclay, University of Amsterdam, Naver Labs, Sorbonne Universite - funded by the European Union |
| Training | 4 trillion tokens across the language set; 400 NVIDIA H100s on the Marenostrum 5 supercomputer; instruction-tuned on EuroBlocks |
| Base model | utter-project/EuroLLM-9B-Instruct (a post-trained utter-project/EuroLLM-9B) |