Quantization made by Richard Erkhov.

Mahou-1.5-mistral-nemo-12B-lorablated - GGUF

Model creator: https://huggingface.co/nbeerbower/
Original model: https://huggingface.co/nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated/

Name	Quant method	Size
Mahou-1.5-mistral-nemo-12B-lorablated.Q2_K.gguf	Q2_K	4.46GB
Mahou-1.5-mistral-nemo-12B-lorablated.IQ3_XS.gguf	IQ3_XS	4.94GB
Mahou-1.5-mistral-nemo-12B-lorablated.IQ3_S.gguf	IQ3_S	5.18GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q3_K_S.gguf	Q3_K_S	5.15GB
Mahou-1.5-mistral-nemo-12B-lorablated.IQ3_M.gguf	IQ3_M	5.33GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q3_K.gguf	Q3_K	5.67GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q3_K_M.gguf	Q3_K_M	5.67GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q3_K_L.gguf	Q3_K_L	6.11GB
Mahou-1.5-mistral-nemo-12B-lorablated.IQ4_XS.gguf	IQ4_XS	6.33GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q4_0.gguf	Q4_0	6.59GB
Mahou-1.5-mistral-nemo-12B-lorablated.IQ4_NL.gguf	IQ4_NL	6.65GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q4_K_S.gguf	Q4_K_S	6.63GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q4_K.gguf	Q4_K	6.96GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q4_K_M.gguf	Q4_K_M	6.96GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q4_1.gguf	Q4_1	7.26GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q5_0.gguf	Q5_0	7.93GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q5_K_S.gguf	Q5_K_S	7.93GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q5_K.gguf	Q5_K	8.13GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q5_K_M.gguf	Q5_K_M	8.13GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q5_1.gguf	Q5_1	8.61GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q6_K.gguf	Q6_K	9.37GB
Mahou-1.5-mistral-nemo-12B-lorablated.Q8_0.gguf	Q8_0	12.13GB

Original model description:

license: apache-2.0 library_name: transformers tags: - mergekit - merge base_model: - flammenai/Mahou-1.5-mistral-nemo-12B - nbeerbower/Mistral-Nemo-12B-abliterated-LORA model-index: - name: Mahou-1.5-mistral-nemo-12B-lorablated results: - task: type: text-generation name: Text Generation dataset: name: IFEval (0-Shot) type: HuggingFaceH4/ifeval args: num_few_shot: 0 metrics: - type: inst_level_strict_acc and prompt_level_strict_acc value: 68.25 name: strict accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: BBH (3-Shot) type: BBH args: num_few_shot: 3 metrics: - type: acc_norm value: 36.08 name: normalized accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MATH Lvl 5 (4-Shot) type: hendrycks/competition_math args: num_few_shot: 4 metrics: - type: exact_match value: 5.29 name: exact match source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: GPQA (0-shot) type: Idavidrein/gpqa args: num_few_shot: 0 metrics: - type: acc_norm value: 3.91 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MuSR (0-shot) type: TAUR-Lab/MuSR args: num_few_shot: 0 metrics: - type: acc_norm value: 16.55 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard - task: type: text-generation name: Text Generation dataset: name: MMLU-PRO (5-shot) type: TIGER-Lab/MMLU-Pro config: main split: test args: num_few_shot: 5 metrics: - type: acc value: 28.6 name: accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Mahou-1.5-mistral-nemo-12B-lorablated name: Open LLM Leaderboard

Mahou-1.5-mistral-nemo-12B-lorablated

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the task arithmetic merge method using flammenai/Mahou-1.5-mistral-nemo-12B + nbeerbower/Mistral-Nemo-12B-abliterated-LORA as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

base_model: flammenai/Mahou-1.5-mistral-nemo-12B+nbeerbower/Mistral-Nemo-12B-abliterated-LORA
dtype: bfloat16
merge_method: task_arithmetic
parameters:
  normalize: false
slices:
- sources:
  - layer_range: [0, 40]
    model: flammenai/Mahou-1.5-mistral-nemo-12B+nbeerbower/Mistral-Nemo-12B-abliterated-LORA
    parameters:
      weight: 1.0

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	26.45
IFEval (0-Shot)	68.25
BBH (3-Shot)	36.08
MATH Lvl 5 (4-Shot)	5.29
GPQA (0-shot)	3.91
MuSR (0-shot)	16.55
MMLU-PRO (5-shot)	28.60

Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more quants, at much higher speed, than I would otherwise be able to.