NeuralDaredevil-SuperNova-Lite-7B-DARETIES-abliterated

NeuralDaredevil-SuperNova-Lite-7B-DARETIES-abliterated is a merge of the following models using LazyMergekit:

• mlabonne/NeuralDaredevil-8B-abliterated
• grimjim/Llama-3.1-SuperNova-Lite-lorabilterated-8B

Quantised versions of this model are available in GGUF format from here Or use the following direct links:

• Q5_K_M
• Q4_K_M

open-llm-leaderboard results

Average	IFEval	BBH	MATH Lvl 5	GPQA	MUSR	MMLU-PRO
27.5	79.99	30.76	10.27	4.14	9.47	30.37	🤗 Open LLM Leaderboard

🧩 Configuration

models:
  - model: NousResearch/Meta-Llama-3.1-8B-Instruct
  - model: mlabonne/NeuralDaredevil-8B-abliterated
    parameters:
      density: 0.53
      weight: 0.55
  - model: grimjim/Llama-3.1-SuperNova-Lite-lorabilterated-8B
    parameters:
      density: 0.53
      weight: 0.45
merge_method: dare_ties
base_model: NousResearch/Meta-Llama-3.1-8B-Instruct
parameters:
  int8_mask: true
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "BoltMonkey/NeuralDaredevil-SuperNova-Lite-7B-DARETIES-ablorabliterated"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])