README.md

metadata

language:
  - en
  - zh
license: apache-2.0
library_name: transformers
tags:
  - mergekit
  - merge
  - custom-research
base_model:
  - sthenno-com/miscii-14b-1028
metrics:
  - accuracy
model-index:
  - name: miscii-14b-1225
    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: 78.78
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          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: 50.91
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          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: 31.57
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          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: 17
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          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: 14.77
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          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: 47.46
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=sthenno-com/miscii-14b-1225
          name: Open LLM Leaderboard

miscii-14b-1225

Image source: Rrhar'il | Phigros

Prompting & Usage

See miscii-14b-1028 for more details.

Training Details

Coming soon

Merge Details

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

Merge Method

This model was merged using the TIES merge method using miscii-14b-1028 as a base.

Models Merged

The following models were included in the merge:

• sthenno/exp-002
• sthenno/miscii-1218

Configuration

The following YAML configuration was used to produce this model:

tokenizer_source: "base"
chat_template: "chatml"

merge_method: ties
dtype: bfloat16

parameters:
  normalize: true

base_model: sthenno-com/miscii-14b-1028

models:
  - model: sthenno-com/miscii-14b-1028
    parameters:
      weight: 1
      density: 0.5
  - model: sthenno/miscii-1218
    parameters:
      weight: 1
      density: 0.5
  - model: sthenno/exp-002
    parameters:
      weight: 0.9
      density: 0.5
  - model: sthenno/miscii-1218
    parameters:
      weight: 0.6
      density: 0.5

Open LLM Leaderboard Evaluation Results

Congratulations to the miscii series models for surpassing 40 points for the first time! As of December 25, 2024, this should be the best-performing 14B model in the tests, right?

Metric	Value
Avg.	40.08
IFEval (0-Shot)	78.78
BBH (3-Shot)	50.91
MATH Lvl 5 (4-Shot)	31.57
GPQA (0-shot)	17.00
MuSR (0-shot)	14.77
MMLU-PRO (5-shot)	47.46