README.md

metadata

license: llama3.3
base_model:
  - allura-forge/Llama-3.3-8B-Instruct
pipeline_tag: text-generation
model-index:
  - name: shb777/Llama-3.3-8B-Instruct-128K
    results:
      - task:
          type: text-generation
        dataset:
          name: BBH
          type: leaderboard
        metrics:
          - type: accuracy
            value: 54.1
            name: acc_norm
      - task:
          type: text-generation
        dataset:
          name: GPQA
          type: leaderboard
        metrics:
          - type: accuracy
            value: 29.9
            name: acc_norm
      - task:
          type: text-generation
        dataset:
          name: MMLU Pro
          type: leaderboard
        metrics:
          - type: accuracy
            value: 38
            name: acc
      - task:
          type: text-generation
        dataset:
          name: MuSR
          type: leaderboard
        metrics:
          - type: accuracy
            value: 37.8
            name: acc_norm
      - task:
          type: text-generation
        dataset:
          name: IFEval
          type: leaderboard
        metrics:
          - type: accuracy
            value: 85.2
            name: avg(prompt_strict + inst_strict)
      - task:
          type: text-generation
        dataset:
          name: MATH Hard
          type: leaderboard
        metrics:
          - type: accuracy
            value: 27.3
            name: exact_match

Llama 3.3 8B 128K Instruct (Fixed)

Original allura-forge/Llama-3.3-8B-Instruct, Thanks!

imatrix GGUF's by mradermacher (Recommended)

static GGUF's

Evals

Additional Fixes:

• Added rope_scaling
• Added chat template (Unsloth) in tokenizer config
• Updated generation config
• Enabled full context length