README.md

metadata

license: apache-2.0
tags:
  - moe
  - frankenmoe
  - merge
  - mergekit
  - lazymergekit
  - aloobun/Cypher-Mini-1.8B
  - aloobun/Cypher-CoT-1.8B
base_model:
  - aloobun/Cypher-Mini-1.8B
  - aloobun/Cypher-CoT-1.8B
model-index:
  - name: Cypher-Mixtral-2x1.8B-v0.1
    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: 40.44
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          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: 67.7
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          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: 31.81
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          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: 39.94
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          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: 65.35
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          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: 10.46
            name: accuracy
        source:
          url: >-
            https://huggingface.co./spaces/HuggingFaceH4/open_llm_leaderboard?query=aloobun/Cypher-Mixtral-2x1.8B-v0.1
          name: Open LLM Leaderboard

Cypher-Laser-Mixtral-2x1.8B-v0.1

Cypher-Laser-Mixtral-2x1.8B-v0.1 is a Mixure of Experts (MoE) made with the following models using LazyMergekit:

• aloobun/Cypher-Mini-1.8B
• aloobun/Cypher-CoT-1.8B

🧩 Configuration

base_model: aloobun/Cypher-Mini-1.8B
gate_mode: hidden
dtype: bfloat16
experts:
  - source_model: aloobun/Cypher-Mini-1.8B
    positive_prompts:
      - "Write a Python script that sorts a list of integers using the bubble sort algorithm."
      - "Write a JavaScript function that redirects a web page to another page after 5 seconds."
      - "Describe the steps to troubleshoot a fluid dynamics issue with a water fountain."
      - "Write a short story about a knight's quest to find a lost treasure, and then summarize it in one paragraph."
      - "Summarize the following article with details and clarity."
      - "Tell me about your favorite book and why you like it."
      
  - source_model: aloobun/Cypher-CoT-1.8B
    positive_prompts:
      - "Liam saw an animal running on the farm. Q: Is it true that The animal could be a horse."
      - "Based on the following paragraph can we conclude that the sentence below is true?"
      - "According to the article, how do dolphins communicate with each other?"
      - "Solve this math problem Solve 7644 = 4648*d - 4557*d for d."
      - "If we have 3 marbles, and two roll under the counter, and one is found, how many marbles are there?"
      - "What is the result of 25 divided by 5?"
      - "Is it morally justifiable to lie to protect someone's feelings?"
      - "Determine if the sentence is true based on the text below. Choose from options."
      - "What might a person do if they forget their umbrella on a rainy day?"
      - "Which of the following is an example of renewable energy: a) Coal, b) Solar, c) Oil, d) Natural gas?"
      - "What is the capital of Canada? a) Toronto, b) Ottawa, c) Montreal, d) Vancouver."
      - "Which of these animals is a mammal? a) Snake, b) Dolphin, c) Turtle, d) Frog."
      - "Given a story, answer the question about the story."
      - "Given a prompt and four completions, select the completion that is the most plausible in continuing or answering the prompt."

Format:

<|system|></s><|prompt|></s><|answer|>

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	42.62
AI2 Reasoning Challenge (25-Shot)	40.44
HellaSwag (10-Shot)	67.70
MMLU (5-Shot)	31.81
TruthfulQA (0-shot)	39.94
Winogrande (5-shot)	65.35
GSM8k (5-shot)	10.46