Instructions to use cs-552-2026-4neurons/general_knowledge_model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cs-552-2026-4neurons/general_knowledge_model with Transformers:

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

pipe = pipeline("text-generation", model="cs-552-2026-4neurons/general_knowledge_model")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cs-552-2026-4neurons/general_knowledge_model")
model = AutoModelForCausalLM.from_pretrained("cs-552-2026-4neurons/general_knowledge_model", 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 cs-552-2026-4neurons/general_knowledge_model with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cs-552-2026-4neurons/general_knowledge_model"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cs-552-2026-4neurons/general_knowledge_model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cs-552-2026-4neurons/general_knowledge_model

SGLang

How to use cs-552-2026-4neurons/general_knowledge_model 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 "cs-552-2026-4neurons/general_knowledge_model" \
    --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": "cs-552-2026-4neurons/general_knowledge_model",
		"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 "cs-552-2026-4neurons/general_knowledge_model" \
        --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": "cs-552-2026-4neurons/general_knowledge_model",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use cs-552-2026-4neurons/general_knowledge_model with Docker Model Runner:
```
docker model run hf.co/cs-552-2026-4neurons/general_knowledge_model
```

general_knowledge_model

Commit History

Automated MNLP evaluation report (2026-06-09) (#12)

f1542ca

benjinoeclara

zechen-nlp commited on Jun 10

Upload folder using huggingface_hub

d3ae2f7
verified

benjinoeclara commited on Jun 5

Upload folder using huggingface_hub

bed7345
verified

benjinoeclara commited on Jun 3

Automated MNLP evaluation report (2026-06-03) (#11)

08fa3b1

benjinoeclara

zechen-nlp commited on Jun 3

Upload folder using huggingface_hub

4601433
verified

benjinoeclara commited on Jun 1

Automated MNLP evaluation report (2026-06-01) (#10)

e9c1c71

benjinoeclara

zechen-nlp commited on Jun 1

Upload folder using huggingface_hub

632402a
verified

benjinoeclara commited on May 31

Upload folder using huggingface_hub

52875a4
verified

benjinoeclara commited on May 29

v8: self-distill 1.7B baseline + tiny LoRA r=8 (preserve native reasoning, format only)

9ecf6b6
verified

benjinoeclara commited on May 29

Automated MNLP evaluation report (2026-05-29) (#9)

08c289a

benjinoeclara

zechen-nlp commited on May 29

v7: distill Qwen3-14B thinking + LoRA r=32 (max_seq=6144, conservative SFT)

c38f8f3
verified

benjinoeclara commited on May 28

Automated MNLP evaluation report (2026-05-25) (#8)

0e86d9e

benjinoeclara

zechen-nlp commited on May 26

v6: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5)

6570318
verified

benjinoeclara commited on May 24

Automated MNLP evaluation report (2026-05-24) (#7)

2f6cefa

benjinoeclara

zechen-nlp commited on May 24

Automated MNLP evaluation report (2026-05-23) (#6)

e612279

benjinoeclara

zechen-nlp commited on May 23

v5_finaly: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5)

8a1522f
verified

benjinoeclara commited on May 23

Automated MNLP evaluation report (2026-05-22) (#5)

1b20bd0

benjinoeclara

zechen-nlp commited on May 22

v6: long-reasoning distill (mmlu/mmlu_world/mmlu_pro_cot prompted for 10-20 sentences). OOD 0.6918 / dev_small 0.6208. Median chars on OOD 606→2634 (4.3x v5). MMLU-style improved (+2.5 to +6.7pt), elementary science (ARC/OBQA) flat-to-negative — interesting trade-off vs v5.

0c3eaaf
verified

benjinoeclara commited on May 22

Automated MNLP evaluation report (2026-05-21) (#4)

bde9d57

benjinoeclara

zechen-nlp commited on May 21

v3: <DESCRIPTION COURTE — ex: distill_cot dataset, pass@1 dev_small=0.XX, OOD=0.YY>

2e63bef
verified

benjinoeclara commited on May 21

Automated MNLP evaluation report (2026-05-20) (#3)

c1da342

benjinoeclara

zechen-nlp commited on May 20

rollback to v1: LoRA SFT baseline (pass@1 ~0.556 on dev_full)

d0c05cf
verified

benjinoeclara commited on May 19

Automated MNLP evaluation report (2026-05-18) (#2)

0e1ecde

benjinoeclara

zechen-nlp commited on May 18

v4: DPO self-improve loop (3.5k pairs from v3, max_length=4096), pass@1 0.579→0.602, 11-20 bucket +4.3pt, humanities +4.5pt

9d8d6e3
verified

benjinoeclara commited on May 18

v3: DPO LoRA on 2.1k pairs (β=0.1, lr=2e-5, 2 ep), pass@1 0.564→0.579, history_geo +5.2pt

6ac3191
verified

benjinoeclara commited on May 18

Automated MNLP evaluation report (2026-05-13) (#1)

ab1454a

benjinoeclara

zechen-nlp commited on May 15

v2: distill MMLU-Pro-CoT+ECQA, 2 epochs, dev set with 11-20

8c752e4
verified

benjinoeclara commited on May 14

v1: LoRA SFT Qwen3-1.7B on GK (pass@1=0.556, pass@8=0.863 on dev_full)

8d141a5
verified

benjinoeclara commited on May 13

Upload config

4635543
verified

nicolasteissier commited on May 11

Upload tokenizer

645ae88
verified

nicolasteissier commited on May 11

Upload Qwen3ForCausalLM

058c6c5
verified

nicolasteissier commited on May 11

initial commit

59ba8b9
verified

nicolasteissier commited on May 11

Commit History

Automated MNLP evaluation report (2026-06-09) (#12) f1542ca

Upload folder using huggingface_hub d3ae2f7 verified

Upload folder using huggingface_hub bed7345 verified

Automated MNLP evaluation report (2026-06-03) (#11) 08fa3b1

Upload folder using huggingface_hub 4601433 verified

Automated MNLP evaluation report (2026-06-01) (#10) e9c1c71

Upload folder using huggingface_hub 632402a verified

Upload folder using huggingface_hub 52875a4 verified

v8: self-distill 1.7B baseline + tiny LoRA r=8 (preserve native reasoning, format only) 9ecf6b6 verified

Automated MNLP evaluation report (2026-05-29) (#9) 08c289a

v7: distill Qwen3-14B thinking + LoRA r=32 (max_seq=6144, conservative SFT) c38f8f3 verified

Automated MNLP evaluation report (2026-05-25) (#8) 0e86d9e

v6: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5) 6570318 verified

Automated MNLP evaluation report (2026-05-24) (#7) 2f6cefa

Automated MNLP evaluation report (2026-05-23) (#6) e612279

v5_finaly: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5) 8a1522f verified

Automated MNLP evaluation report (2026-05-22) (#5) 1b20bd0

v6: long-reasoning distill (mmlu/mmlu_world/mmlu_pro_cot prompted for 10-20 sentences). OOD 0.6918 / dev_small 0.6208. Median chars on OOD 606→2634 (4.3x v5). MMLU-style improved (+2.5 to +6.7pt), elementary science (ARC/OBQA) flat-to-negative — interesting trade-off vs v5. 0c3eaaf verified

Automated MNLP evaluation report (2026-05-21) (#4) bde9d57

v3: <DESCRIPTION COURTE — ex: distill_cot dataset, pass@1 dev_small=0.XX, OOD=0.YY> 2e63bef verified

Automated MNLP evaluation report (2026-05-20) (#3) c1da342

rollback to v1: LoRA SFT baseline (pass@1 ~0.556 on dev_full) d0c05cf verified

Automated MNLP evaluation report (2026-05-18) (#2) 0e1ecde

v4: DPO self-improve loop (3.5k pairs from v3, max_length=4096), pass@1 0.579→0.602, 11-20 bucket +4.3pt, humanities +4.5pt 9d8d6e3 verified

v3: DPO LoRA on 2.1k pairs (β=0.1, lr=2e-5, 2 ep), pass@1 0.564→0.579, history_geo +5.2pt 6ac3191 verified

Automated MNLP evaluation report (2026-05-13) (#1) ab1454a

v2: distill MMLU-Pro-CoT+ECQA, 2 epochs, dev set with 11-20 8c752e4 verified

v1: LoRA SFT Qwen3-1.7B on GK (pass@1=0.556, pass@8=0.863 on dev_full) 8d141a5 verified

Upload config 4635543 verified

Upload tokenizer 645ae88 verified

Upload Qwen3ForCausalLM 058c6c5 verified

initial commit 59ba8b9 verified

Automated MNLP evaluation report (2026-06-09) (#12)

f1542ca

Upload folder using huggingface_hub

d3ae2f7
verified

Upload folder using huggingface_hub

bed7345
verified

Automated MNLP evaluation report (2026-06-03) (#11)

08fa3b1

Upload folder using huggingface_hub

4601433
verified

Automated MNLP evaluation report (2026-06-01) (#10)

e9c1c71

Upload folder using huggingface_hub

632402a
verified

Upload folder using huggingface_hub

52875a4
verified

v8: self-distill 1.7B baseline + tiny LoRA r=8 (preserve native reasoning, format only)

9ecf6b6
verified

Automated MNLP evaluation report (2026-05-29) (#9)

08c289a

v7: distill Qwen3-14B thinking + LoRA r=32 (max_seq=6144, conservative SFT)

c38f8f3
verified

Automated MNLP evaluation report (2026-05-25) (#8)

0e86d9e

v6: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5)

6570318
verified

Automated MNLP evaluation report (2026-05-24) (#7)

2f6cefa

Automated MNLP evaluation report (2026-05-23) (#6)

e612279

v5_finaly: long-reasoning distillation, OOD 0.6918, dev_small 0.6208 (median 2634 chars vs 606 v5)

8a1522f
verified

Automated MNLP evaluation report (2026-05-22) (#5)

1b20bd0

v6: long-reasoning distill (mmlu/mmlu_world/mmlu_pro_cot prompted for 10-20 sentences). OOD 0.6918 / dev_small 0.6208. Median chars on OOD 606→2634 (4.3x v5). MMLU-style improved (+2.5 to +6.7pt), elementary science (ARC/OBQA) flat-to-negative — interesting trade-off vs v5.

0c3eaaf
verified

Automated MNLP evaluation report (2026-05-21) (#4)

bde9d57

v3: <DESCRIPTION COURTE — ex: distill_cot dataset, pass@1 dev_small=0.XX, OOD=0.YY>

2e63bef
verified

Automated MNLP evaluation report (2026-05-20) (#3)

c1da342

rollback to v1: LoRA SFT baseline (pass@1 ~0.556 on dev_full)

d0c05cf
verified

Automated MNLP evaluation report (2026-05-18) (#2)

0e1ecde

v4: DPO self-improve loop (3.5k pairs from v3, max_length=4096), pass@1 0.579→0.602, 11-20 bucket +4.3pt, humanities +4.5pt

9d8d6e3
verified

v3: DPO LoRA on 2.1k pairs (β=0.1, lr=2e-5, 2 ep), pass@1 0.564→0.579, history_geo +5.2pt

6ac3191
verified

Automated MNLP evaluation report (2026-05-13) (#1)

ab1454a

v2: distill MMLU-Pro-CoT+ECQA, 2 epochs, dev set with 11-20

8c752e4
verified

v1: LoRA SFT Qwen3-1.7B on GK (pass@1=0.556, pass@8=0.863 on dev_full)

8d141a5
verified

Upload config

4635543
verified

Upload tokenizer

645ae88
verified

Upload Qwen3ForCausalLM

058c6c5
verified

initial commit

59ba8b9
verified