Instructions to use ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct

SGLang

How to use ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct with Docker Model Runner:
```
docker model run hf.co/ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct
```

Control-LLM-Llama3.1-8B-Math16

This is a fine-tuned model of Llama-3.1-8B-Instruct for mathematical tasks on OpenMath2 dataset.

Linked Paper

This model is associated with the paper: Control-LLM.

Linked Open Source code - training, eval and benchmark

This model is associated with the github: Control-LLM.

Evaluation Results

Here is an overview of the evaluation results and findings:

Benchmark Results Table

The table below summarizes evaluation results across mathematical tasks and original capabilities.

Model	MH	M	G8K	M-Avg	ARC	GPQA	MLU	MLUP	O-Avg	Overall
Llama3.1-8B-Inst	23.7	50.9	85.6	52.1	83.4	29.9	72.4	46.7	60.5	56.3
Control LLM*	36.0	61.7	89.7	62.5	82.5	30.8	71.6	45.4	57.6	60.0

Explanation:

MH: MathHard
M: Math
G8K: GSM8K
M-Avg: Math - Average across MathHard, Math, and GSM8K
ARC: ARC benchmark
GPQA: General knowledge QA
MLU: MMLU (Massive Multitask Language Understanding)
MLUP: MMLU Pro
O-Avg: Original Capability - Average across ARC, GPQA, MMLU, and MLUP
Overall: Combined average across all tasks

Catastrophic Forgetting on OpenMath

The following plot illustrates and compares catastrophic forgetting mitigation during training

Alignment Result

The plot below highlights the alignment result of the model trained with Control LLM.

Downloads last month: 15

Safetensors

Model size

10B params

Tensor type

BF16

Model tree for ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct

Base model

meta-llama/Llama-3.1-8B

Finetuned

meta-llama/Llama-3.1-8B-Instruct

Finetuned

(2733)

this model

Dataset used to train ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct

Paper for ControlLLM/Control-LLM-Llama3.1-8B-Math16-Instruct

Control LLM: Controlled Evolution for Intelligence Retention in LLM

Paper • 2501.10979 • Published Jan 19, 2025 • 6

Evaluation results

exact_match,none on Math, Math Hard, GSM8K
self-reported

0.621
exact_match,none (gsm8k_0shot_instruct) on Math, Math Hard, GSM8K
self-reported

0.897
exact_match,none (meta_math_0shot_instruct) on Math, Math Hard, GSM8K
self-reported

0.617
exact_match,none (meta_math_hard_0shot_instruct) on Math, Math Hard, GSM8K
self-reported

0.360
exact_match,strict-match on Llama-3.1-8B-Instruct-evals Dataset
self-reported

0.600
exact_match,strict-match (meta_arc_0shot_instruct) on Llama-3.1-8B-Instruct-evals Dataset
self-reported

0.825
exact_match,strict-match (meta_gpqa_0shot_cot_instruct) on Llama-3.1-8B-Instruct-evals Dataset
self-reported

0.308
exact_match,strict-match (meta_mmlu_0shot_instruct) on Llama-3.1-8B-Instruct-evals Dataset
self-reported

0.716
exact_match,strict-match (meta_mmlu_pro_5shot_instruct) on Llama-3.1-8B-Instruct-evals Dataset
self-reported

0.454