README.md

metadata

language:
  - en
license: llama2
tags:
  - roleplay
  - conversational
datasets:
  - migtissera/Synthia-v1.3
  - open-llm-leaderboard/details_NousResearch__Nous-Hermes-llama-2-7b
pipeline_tag: text-generation
model-index:
  - name: L2-7b-Hermes-Synthia
    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: 51.02
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          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: 79.12
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          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: 47.88
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          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: 46.77
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          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: 74.51
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          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: 13.95
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=LTC-AI-Labs/L2-7b-Hermes-Synthia
          name: Open LLM Leaderboard

Fine-tuned the synthia dataset on the hermes2 7b model

In my opinion it's probably the best model I fine-tuned in-terms of role-playing (tested on LavernAI)

Future plans:

• I'll probably do more test in other areas

• Will add other languages (Potentially japanese and chinese)

• Finetune it on mistral models?

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	52.21
AI2 Reasoning Challenge (25-Shot)	51.02
HellaSwag (10-Shot)	79.12
MMLU (5-Shot)	47.88
TruthfulQA (0-shot)	46.77
Winogrande (5-shot)	74.51
GSM8k (5-shot)	13.95