metadata

license: llama3.1
datasets:
  - DebateLabKIT/deepa2-conversations
  - DebateLabKIT/deep-argmap-conversations
  - allenai/tulu-3-sft-mixture
base_model:
  - meta-llama/Llama-3.1-8B-Instruct
pipeline_tag: text-generation
library_name: transformers
tags:
  - logic
  - argumentation
  - critical-thinking
  - argument-mapping
  - trl
  - sft
model-index:
  - name: Llama-3.1-Argunaut-1-8B-SFT
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: IFEval (0-Shot)
          type: wis-k/instruction-following-eval
          split: train
          args:
            num_few_shot: 0
        metrics:
          - type: inst_level_strict_acc and prompt_level_strict_acc
            value: 55.19
            name: averaged accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: BBH (3-Shot)
          type: SaylorTwift/bbh
          split: test
          args:
            num_few_shot: 3
        metrics:
          - type: acc_norm
            value: 27.19
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MATH Lvl 5 (4-Shot)
          type: lighteval/MATH-Hard
          split: test
          args:
            num_few_shot: 4
        metrics:
          - type: exact_match
            value: 11.18
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GPQA (0-shot)
          type: Idavidrein/gpqa
          split: train
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 4.47
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          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: 15.85
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          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: 27.47
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard#/?search=DebateLabKIT%2FLlama-3.1-Argunaut-1-8B-SFT
          name: Open LLM Leaderboard

Model Card for Llama-3.1-Argunaut-1-8B-SFT

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct. It has been trained using TRL.

Quick start

from transformers import pipeline

question = "Are you familiar with Argdown syntax? What's its purpose?"
generator = pipeline("text-generation", model="DebateLabKIT/Llama-3.1-Argunaut-1-8B-SFT", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Evals

⚠️ NOTE: These self-reported results have been obtained with lm-eval-harness and using local-completions api; they deviate significantly from the official Open LLM Leaderboard evals, which are also reported at the end of this readme.

LM Eval Harness results (local completions/vllm): wandb report

Model	BBH	MATH	GPQA	MMLU Pro
Llama-3.1-Argunaut-1-8B-SFT	44.6%	9.0%	32.1%	34.5%

SFT dataset mixture

Dataset	Weight (examples)	Weight (tokens)
DebateLabKIT/deepa2-conversations	25%	49%
DebateLabKIT/deep-argmap-conversations	25%	18%
allenai/tulu-3-sft-mixture	50%	33%

Training procedure

Trained with SFT on 1M examples and for 1 epoch with

context length 8196
packing (trl implementation)
spectrum (top 30 percent)

# Training parameters
num_train_epochs: 1
per_device_train_batch_size: 8
gradient_accumulation_steps: 2
gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
learning_rate: 5.0e-6  # following _Tülu 3_ recipe
lr_scheduler_type: cosine
warmup_ratio: 0.1

Hardware: 2 x H100 GPUs.

This work was performed on the HoreKa supercomputer funded by the Ministry of Science, Research and the Arts Baden-Württemberg and by the Federal Ministry of Education and Research.

Framework versions

TRL: 0.12.1
Transformers: 4.46.3
Pytorch: 2.4.1
Datasets: 3.1.0
Tokenizers: 0.20.3

Credits

This work wouldn't be possible without all the great contributions from the open LLM community. Thank you! Special kudos go to

@philschmid for his latest fine-tuning boilerplate
@lvwerra, @lewtun et al for building and maintaining trl
@cognitivecomputations for sharing spectrum

Open LLM Leaderboard Evaluation Results

Detailed results can be found here! Summarized results can be found here!

Metric	Value (%)
Average	23.56
IFEval (0-Shot)	55.19
BBH (3-Shot)	27.19
MATH Lvl 5 (4-Shot)	11.18
GPQA (0-shot)	4.47
MuSR (0-shot)	15.85
MMLU-PRO (5-shot)	27.47