Dampfinchen's picture
Adding Evaluation Results (#1)
4818ea2 verified
metadata
license: llama3
library_name: transformers
tags:
  - mergekit
  - merge
base_model:
  - Dampfinchen/Llama-3-8B-Ultra-Instruct
  - NousResearch/Meta-Llama-3-8B
  - NousResearch/Meta-Llama-3-8B-Instruct
model-index:
  - name: Llama-3-8B-Ultra-Instruct-SaltSprinkle
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            value: 61.35
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            value: 77.76
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 67.88
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 52.82
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 74.98
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 70.89
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Dampfinchen/Llama-3-8B-Ultra-Instruct-SaltSprinkle
          name: Open LLM Leaderboard

merge

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using NousResearch/Meta-Llama-3-8B as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: NousResearch/Meta-Llama-3-8B-Instruct
    parameters:
      density: 1
      weight: 1
  - model: Dampfinchen/Llama-3-8B-Ultra-Instruct
    parameters:
      density: 0.5
      weight: 0.2

merge_method: dare_ties
base_model: NousResearch/Meta-Llama-3-8B
dtype: bfloat16

Test of salt sprinkle methode. The goal is to retain all of L3 Instruct's capabilities while adding better RP, RAG, German and story writing capabilities in the form of Ultra Instruct. Model may generate harmful responses, I'm not responsible for what you do with this model.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric Value
Avg. 67.61
AI2 Reasoning Challenge (25-Shot) 61.35
HellaSwag (10-Shot) 77.76
MMLU (5-Shot) 67.88
TruthfulQA (0-shot) 52.82
Winogrande (5-shot) 74.98
GSM8k (5-shot) 70.89