Instructions to use SlowGuess/ABForge-Qwen3-8B-Combined with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SlowGuess/ABForge-Qwen3-8B-Combined with Transformers:

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

pipe = pipeline("text-generation", model="SlowGuess/ABForge-Qwen3-8B-Combined")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SlowGuess/ABForge-Qwen3-8B-Combined")
model = AutoModelForCausalLM.from_pretrained("SlowGuess/ABForge-Qwen3-8B-Combined", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use SlowGuess/ABForge-Qwen3-8B-Combined with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "SlowGuess/ABForge-Qwen3-8B-Combined"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SlowGuess/ABForge-Qwen3-8B-Combined",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/SlowGuess/ABForge-Qwen3-8B-Combined

SGLang

How to use SlowGuess/ABForge-Qwen3-8B-Combined 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 "SlowGuess/ABForge-Qwen3-8B-Combined" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SlowGuess/ABForge-Qwen3-8B-Combined",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "SlowGuess/ABForge-Qwen3-8B-Combined" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "SlowGuess/ABForge-Qwen3-8B-Combined",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use SlowGuess/ABForge-Qwen3-8B-Combined with Docker Model Runner:
```
docker model run hf.co/SlowGuess/ABForge-Qwen3-8B-Combined
```

ABForge-Qwen3-8B-Combined

The main ABForge model: a single unified checkpoint that performs both tasks of paper-grounded ablation design.

ABForge is a post-training pipeline for paper-grounded ablation design. This checkpoint is post-trained with the full ABForge pipeline on a 1:1 mixture of both tasks — supervised fine-tuning from Qwen/Qwen3-8B followed by rubric-guided GRPO with per-task reward routing (SFT → GRPO, RL step 200).

Tasks

Given the ablation-free context of a research paper, this one model handles both:

Task 1: Ablation Objective Generation — propose candidate ablation objectives, each expressed as a Target Module (the component to ablate) paired with a Research Question it is meant to answer.
Task 2: Ablation Experiment Design — produce a concrete, executable ablation experiment plan (variants, controls, datasets, metrics, expected outcomes) for a given objective.

Training data

Mixed-task SFT on train/sft_task1_45961.jsonl + train/sft_task2_37019.jsonl, then mixed-task GRPO on train/RL_task1_30K.jsonl + train/RL_task2_30K.jsonl, from SlowGuess/abforge-data (derived from CC-licensed research papers). Both stages use a 1:1 task mixture, and during RL each rollout is routed to its task-specific reward by data_source. Evaluation uses the held-out AblationBench split of the same dataset (eval/ablationbench_1000.jsonl).

Results

Full AblationBench, automated rubric-based LLM-as-a-Judge evaluation:

Model	Task 1 (%)	Task 2 (%)
`Qwen/Qwen3-8B` (base)	44.4	43.4
ABForge-Qwen3-8B-Combined (this model)	55.9	62.4

The unified model surpasses the task-specific specialist on Task 2 while consolidating both capabilities into a single checkpoint; see the paper's ablation table for the task-specific comparison.

Related models

Task-specific specialists (ablation of task sharing):

SlowGuess/ABForge-Qwen3-8B-Task1 (SFT · RL-only)
SlowGuess/ABForge-Qwen3-8B-Task2 (SFT · RL-only)

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "SlowGuess/ABForge-Qwen3-8B-Combined"
tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto", device_map="auto")

Use the Task 1 / Task 2 prompt templates from the code release — the model is trained on those exact formats and the rubric-based evaluator expects the corresponding output structure.

Evaluation

Reproduce AblationBench evaluation with the SlowGuess/Abforge_1 code:

git clone https://github.com/SlowGuess/Abforge_1 && cd Abforge_1
huggingface-cli download SlowGuess/abforge-data --repo-type dataset --local-dir data

Downloads last month: 28

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for SlowGuess/ABForge-Qwen3-8B-Combined

Base model

Qwen/Qwen3-8B-Base

Finetuned

Qwen/Qwen3-8B

Finetuned

(1997)

this model