Advancing Open-source Language Models with Mixed-Quality Data

with 32k context

Usage

To use this model, we highly recommend installing the OpenChat package by following the installation guide in our repository and using the OpenChat OpenAI-compatible API server by running the serving command from the table below. The server is optimized for high-throughput deployment using vLLM and can run on a consumer GPU with 24GB RAM. To enable tensor parallelism, append --tensor-parallel-size N to the serving command.

Once started, the server listens at localhost:18888 for requests and is compatible with the OpenAI ChatCompletion API specifications. Please refer to the example request below for reference. Additionally, you can use the OpenChat Web UI for a user-friendly experience.

If you want to deploy the server as an online service, you can use --api-keys sk-KEY1 sk-KEY2 ... to specify allowed API keys and --disable-log-requests --disable-log-stats --log-file openchat.log for logging only to a file. For security purposes, we recommend using an HTTPS gateway in front of the server.

Model	Size	Context	Weights	Serving
OpenChat-3.5-0106	7B	8192	Huggingface	`python -m ochat.serving.openai_api_server --model openchat/openchat-3.5-0106 --engine-use-ray --worker-use-ray`

Example request (click to expand)

💡 Default Mode (GPT4 Correct): Best for coding, chat and general tasks

curl http://localhost:18888/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openchat_3.5",
    "messages": [{"role": "user", "content": "You are a large language model named OpenChat. Write a poem to describe yourself"}]
  }'

🧮 Mathematical Reasoning Mode: Tailored for solving math problems

curl http://localhost:18888/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openchat_3.5",
    "condition": "Math Correct",
    "messages": [{"role": "user", "content": "10.3 − 7988.8133 = "}]
  }'

Conversation templates

💡 Default Mode (GPT4 Correct): Best for coding, chat and general tasks

GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi<|end_of_turn|>GPT4 Correct User: How are you today?<|end_of_turn|>GPT4 Correct Assistant:

🧮 Mathematical Reasoning Mode: Tailored for solving math problems

Math Correct User: 10.3 − 7988.8133=<|end_of_turn|>Math Correct Assistant:

⚠️ Notice: Remember to set <|end_of_turn|> as end of generation token.

The default (GPT4 Correct) template is also available as the integrated tokenizer.chat_template, which can be used instead of manually specifying the template:

messages = [
    {"role": "user", "content": "Hello"},
    {"role": "assistant", "content": "Hi"},
    {"role": "user", "content": "How are you today?"}
]
tokens = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
assert tokens == [1, 420, 6316, 28781, 3198, 3123, 1247, 28747, 22557, 32000, 420, 6316, 28781, 3198, 3123, 21631, 28747, 15359, 32000, 420, 6316, 28781, 3198, 3123, 1247, 28747, 1602, 460, 368, 3154, 28804, 32000, 420, 6316, 28781, 3198, 3123, 21631, 28747]

(Experimental) Evaluator / Feedback Capabilities

We've included evaluator capabilities in this release to advance open-source models as evaluators. You can use Default Mode (GPT4 Correct) with the following prompt (same as Prometheus) to evaluate a response.

###Task Description:
An instruction (might include an Input inside it), a response to evaluate, a reference answer that gets a score of 5, and a score rubric representing a evaluation criteria are given.
1. Write a detailed feedback that assess the quality of the response strictly based on the given score rubric, not evaluating in general.
2. After writing a feedback, write a score that is an integer between 1 and 5. You should refer to the score rubric.
3. The output format should look as follows: "Feedback: (write a feedback for criteria) [RESULT] (an integer number between 1 and 5)"
4. Please do not generate any other opening, closing, and explanations.

###The instruction to evaluate:
{orig_instruction}

###Response to evaluate:
{orig_response}

###Reference Answer (Score 5):
{orig_reference_answer}

###Score Rubrics:
[{orig_criteria}]
Score 1: {orig_score1_description}
Score 2: {orig_score2_description}
Score 3: {orig_score3_description}
Score 4: {orig_score4_description}
Score 5: {orig_score5_description}

###Feedback:

Benchmarks

Model	# Params	Average	MT-Bench	HumanEval	BBH MC	AGIEval	TruthfulQA	MMLU	GSM8K	BBH CoT
OpenChat-3.5-0106	7B	64.5	7.8	71.3	51.5	49.1	61.0	65.8	77.4	62.2
OpenChat-3.5-1210	7B	63.8	7.76	68.9	49.5	48.0	61.8	65.3	77.3	61.8
OpenChat-3.5	7B	61.6	7.81	55.5	47.6	47.4	59.1	64.3	77.3	63.5
ChatGPT (March)*	???B	61.5	7.94	48.1	47.6	47.1	57.7	67.3	74.9	70.1

OpenHermes 2.5	7B	59.3	7.54	48.2	49.4	46.5	57.5	63.8	73.5	59.9
OpenOrca Mistral	7B	52.7	6.86	38.4	49.4	42.9	45.9	59.3	59.1	58.1
Zephyr-β^	7B	34.6	7.34	22.0	40.6	39.0	40.8	39.8	5.1	16.0
Mistral	7B	-	6.84	30.5	39.0	38.0	-	60.1	52.2	-

Evaluation Details(click to expand)

*: ChatGPT (March) results are from GPT-4 Technical Report, Chain-of-Thought Hub, and our evaluation. Please note that ChatGPT is not a fixed baseline and evolves rapidly over time.

^: Zephyr-β often fails to follow few-shot CoT instructions, likely because it was aligned with only chat data but not trained on few-shot data.

**: Mistral and Open-source SOTA results are taken from reported results in instruction-tuned model papers and official repositories.

All models are evaluated in chat mode (e.g. with the respective conversation template applied). All zero-shot benchmarks follow the same setting as in the AGIEval paper and Orca paper. CoT tasks use the same configuration as Chain-of-Thought Hub, HumanEval is evaluated with EvalPlus, and MT-bench is run using FastChat. To reproduce our results, follow the instructions in our repository.

HumanEval+

Model	Size	HumanEval+ pass@1
OpenChat-3.5-0106	7B	65.9
ChatGPT (December 12, 2023)	???B	64.6
WizardCoder-Python-34B-V1.0	34B	64.6
OpenChat 3.5 1210	7B	63.4
OpenHermes 2.5	7B	41.5

OpenChat-3.5 vs. Grok

🔥 OpenChat-3.5-0106 (7B) now outperforms Grok-0 (33B) on all 4 benchmarks and Grok-1 (???B) on average and 3/4 benchmarks.

	License	# Param	Average	MMLU	HumanEval	MATH	GSM8k
OpenChat-3.5-0106	Apache-2.0	7B	61.0	65.8	71.3	29.3	77.4
OpenChat-3.5-1210	Apache-2.0	7B	60.1	65.3	68.9	28.9	77.3
OpenChat-3.5	Apache-2.0	7B	56.4	64.3	55.5	28.6	77.3
Grok-0	Proprietary	33B	44.5	65.7	39.7	15.7	56.8
Grok-1	Proprietary	???B	55.8	73	63.2	23.9	62.9

*: Grok results are reported by X.AI.

Limitations

Foundation Model Limitations Despite its advanced capabilities, OpenChat is still bound by the limitations inherent in its foundation models. These limitations may impact the model's performance in areas such as:

Complex reasoning
Mathematical and arithmetic tasks
Programming and coding challenges

Hallucination of Non-existent Information OpenChat may sometimes generate information that does not exist or is not accurate, also known as "hallucination". Users should be aware of this possibility and verify any critical information obtained from the model.

Safety OpenChat may sometimes generate harmful, hate speech, biased responses, or answer unsafe questions. It's crucial to apply additional AI safety measures in use cases that require safe and moderated responses.

License

Our OpenChat 3.5 code and models are distributed under the Apache License 2.0.

Citation

@article{wang2023openchat,
  title={OpenChat: Advancing Open-source Language Models with Mixed-Quality Data},
  author={Wang, Guan and Cheng, Sijie and Zhan, Xianyuan and Li, Xiangang and Song, Sen and Liu, Yang},
  journal={arXiv preprint arXiv:2309.11235},
  year={2023}
}

💌 Main Contributor

Wang Guan [[email protected]], Cheng Sijie [[email protected]], Alpay Ariyak [[email protected]]
We look forward to hearing you and collaborating on this exciting project!

LoneStriker
/

openchat-3.5-0106-11b-3.0bpw-h6-exl2