Gradio

INTRO TEXT TEST


🟢	ERNIE-3.5-8K	50.3	51.8	61.3	66.1	44.3	26.9	53.6


🟢	ERNIE-3.5-8K	50.3	51.8	61.3	66.1	44.3	26.9	53.6
🟢	ChatGPT	51.3	46.1	59.9	61.7	45.8	15.4	52
🟢	Glm-4	44.5	45.9	54.4	57.8	40.2	20.3	47.5
🟢	InternLLM2	37.7	34.7	45.1	42.5	37.8	0.7	38.7
🟢	Qwen2	36.7	26	36.8	32.9	36.3	0.2	33.6
🔶	Deepseek-R1	35.5	24.4	34.1	32.1	33.8	0	31.9
🔶	SecGPT	35.7	13.2	31.1	24.2	33	0	28
🟢	Gemma	25.5	17.5	32.5	26.3	25.4	0	25
🟢	Yi	24.4	20.2	29.1	28	22.6	2.1	24.4
🟢	Llama3	31.2	14.9	22.2	19.8	29.5	0	24.2
🟢	Mistral-v0.3	23	14.1	28.5	23.5	21.8	0.1	21.9
🔶	Atom	27.7	9.5	16.6	13.5	26.1	0	19.6
🔶	AutoAudit	18.6	15.3	16.4	17	18.2	0	17.1
🟢	Baichuan2	19.1	13.3	16.8	17	17.8	0	16.8

Some good practices before submitting a model

1) Make sure you can load your model and tokenizer using AutoClasses:

from transformers import AutoConfig, AutoModel, AutoTokenizer
config = AutoConfig.from_pretrained("your model name", revision=revision)
model = AutoModel.from_pretrained("your model name", revision=revision)
tokenizer = AutoTokenizer.from_pretrained("your model name", revision=revision)

If this step fails, follow the error messages to debug your model before submitting it. It's likely your model has been improperly uploaded.

Note: make sure your model is public! Note: if your model needs use_remote_code=True, we do not support this option yet but we are working on adding it, stay posted!

2) Convert your model weights to safetensors

It's a new format for storing weights which is safer and faster to load and use. It will also allow us to add the number of parameters of your model to the Extended Viewer!

3) Make sure your model has an open license!

This is a leaderboard for Open LLMs, and we'd love for as many people as possible to know they can use your model 🤗

4) Fill up your model card

When we add extra information about models to the leaderboard, it will be automatically taken from the model card

In case of model failure

If your model is displayed in the FAILED category, its execution stopped. Make sure you have followed the above steps first. If everything is done, check you can launch the EleutherAIHarness on your model locally, using the above command without modifications (you can add --limit to limit the number of examples per task).


🟢	ERNIE-3.5-8K	10.9	1.7	10.7	72.4	40	12.5	22.9	72.3	83.4	83.4	70.2	75.9	72.8	75.2	65	65.7	67.7	69	26.9	39.7	20.8	35	76	85.6	86.1	76.8	76.2	56.9	62.7	23.4	22.9	50.3	51.8	61.3	66.1	44.3	26.9	53.6	pretrained	?	?	true
🟢	ChatGPT	14.3	2.8	14	75.6	42.1	13.8	24.2	73	86.5	86.5	69.4	72.8	70.8	72.4	56.5	58.9	47.3	48.9	15.4	45.2	24.9	40.5	77.3	72.6	74.9	62.9	66.8	78.1	80.8	20.8	22.5	51.3	46.1	59.9	61.7	45.8	15.4	52	pretrained	?	?	true
🟢	Glm-4	6.1	0.9	6.1	53.9	44.1	16.4	26.3	75.4	77.4	77.2	54.1	64.2	58.6	62.9	48.5	53.6	57.9	59.3	20.3	39.7	22.4	36	75.4	64.3	70.4	74.8	67.4	77.1	78.6	1.6	0.6	44.5	45.9	54.4	57.8	40.2	20.3	47.5	pretrained	float16	?	true	main
🟢	InternLLM2	11.9	1.8	11.7	74.1	42.5	14.6	26.1	75.3	15.1	11.2	56.9	63.9	60	63.1	50	54.1	51	53	0.7	17.6	7.3	15.4	62.9	34.6	47.7	67.6	64.9	71.9	71.9	1	1.3	37.7	34.7	45.1	42.5	37.8	0.7	38.7	pretrained	float16	?	true	main
🟢	Qwen2	11	2.2	10.9	72.8	43.9	16.3	27.5	76.2	25	10.5	50.2	57.2	53.3	56.4	53.5	57.4	7.3	12.3	0.2	17.1	6.7	15	64.5	26.6	35.2	29.4	40.5	80.4	78	2.1	2.6	36.7	26	36.8	32.9	36.3	0.2	33.6	pretrained	float16	?	true	main
🔶	Deepseek-R1	6.1	1.3	6.1	65.6	38.5	11.2	22.2	73.4	24.7	9.8	56.6	61.5	58.8	60.9	60.5	54.3	14.4	12.6	0	8.3	2.5	7	60.2	20.3	21.3	40.2	47.6	72.8	71	0	0	35.5	24.4	34.1	32.1	33.8	0	31.9	fine-tuned	float16	?	true	main
🔶	SecGPT	3.4	0.6	3.4	60.9	39.6	13.3	25.2	73.7	24.7	9.8	59.5	62.4	60.8	62.1	16.5	25.7	6	7.1	0	6	1.5	5.4	50.7	14.4	11.4	29.2	38.7	78.8	74.5	1	1	35.7	13.2	31.1	24.2	33	0	28	fine-tuned	float16	?	true	main
🟢	Gemma	2.7	0.4	2.7	58.7	40.6	14.1	26.1	74.7	16	9.1	27.1	28.6	27.8	28.4	24	32.5	11.2	15.5	0	8.3	2.3	7.4	56.7	35.9	46.6	5.6	10.3	83.7	76.8	0.3	0.4	25.5	17.5	32.5	26.3	25.4	0	25	pretrained	float16	?	true	main
🟢	Yi	3.2	0.5	3.2	58.7	14.7	4.7	10	40.1	24.7	9.8	41.8	43.8	42.7	43.6	53.5	58.3	4.8	8	2.1	6.4	1.9	5.6	41.4	20.5	18.3	13.4	22.1	81.6	76.7	0.1	0.2	24.4	20.2	29.1	28	22.6	2.1	24.4	pretrained	float16	?	true	main
🟢	Llama3	2.2	0.3	2.2	56.5	38	12.7	24.7	71.9	21.6	9.5	47.5	50.9	49	50.5	19	27.5	10.4	11	0	6	1.5	5.3	53.5	24.3	30.1	12.3	20.6	36.6	49.3	1.9	2.5	31.2	14.9	22.2	19.8	29.5	0	24.2	pretrained	float16	?	true	main
🟢	Mistral-v0.3	3.8	0.5	3.8	61.9	39.5	13.4	25.6	71.3	24.7	9.8	16.1	17.7	16.8	17.5	28	35.2	1.2	2.2	0.1	5.4	1.2	4.8	49.1	18	23.3	18.2	28.2	66.7	67.5	4	2.5	23	14.1	28.5	23.5	21.8	0.1	21.9	pretrained	float16	?	true	main
🔶	Atom	3.3	0.4	3.3	57.3	12.6	3.7	9.4	27.9	22	9.6	59.1	60.1	59.4	60	9	14.8	0.2	0.3	0	5.4	1.4	4.7	49.8	3.6	6.2	2.1	3.9	54.4	62.4	0.4	0	27.7	9.5	16.6	13.5	26.1	0	19.6	fine-tuned	float16	?	true	main
🔶	AutoAudit	3.9	0.5	3.9	61	17.9	5.4	13.7	56.8	24.7	9.8	15.2	15.8	15.5	15.7	29	36.8	2.5	4.1	0	6.9	2	6.3	50.1	7.4	11.9	12.4	20.3	32.6	46.7	0	0	18.6	15.3	16.4	17	18.2	0	17.1	fine-tuned	float16	?	true	main
🟢	Baichuan2	3	0.4	3	57.7	22.9	6.8	16	47.6	24.7	9.8	18.2	19.4	18.7	19.2	30.5	37.9	0	0	0	3.9	1	3.5	43	8.1	13.9	3.4	6.4	43.7	58.7	0.3	0.2	19.1	13.3	16.8	17	17.8	0	16.8	pretrained	float16	?	true	main