Dolphin 2.8 Experiment26 7B By cognitivecomputations: Benchmarks, Features and Detailed Analysis. Insights on Dolphin 2.8 Experiment26 7B.

Autotrain compatible Conversational Dataset:ehartford/dolphin Dataset:ehartford/dolphin-code... Dataset:jondurbin/airoboros-2.... Dataset:m-a-p/code-feedback Dataset:teknium/openhermes En Endpoints compatible Mistral Pytorch Region:us Sharded

Model Card on HF 🤗: https://huggingface.co/cognitivecomputations/dolphin-2.8-experiment26-7b

Dolphin 2.8 Experiment26 7B Benchmarks

ARC: 63.65 vs 96.7 (so35)^-34.2%

HellaSwag: 83.7 vs 95.3 (gpt4)^-12.2%

MMLU: 62.31 vs 88.3 (so35)^-29.4%

TruthfulQA: 55.1 vs 59 (gpt4)^-6.6%

WinoGrande: 78.77 vs 87.5 (gpt4)^-10%

GSM8K: 62.62 vs 96.4 (so35)^-35%

LLME Score: 0.19072

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Dolphin 2.8 Experiment26 7B (cognitivecomputations/dolphin-2.8-experiment26-7b)

Dolphin 2.8 Experiment26 7B Parameters and Internals

Model Type

text generation, coding

Use Cases

Areas:

coding, chatbot

Primary Use Cases:

Coding, General chat use-cases, Structured output, Agent cases

Additional Notes

Sponsored by MassedCompute and made possible by community and organization contributions. Uses ChatML format for prompts.

Training Details

Data Sources:

ehartford/dolphin, jondurbin/airoboros-2.2.1, ehartford/dolphin-coder, teknium/openhermes, m-a-p/Code-Feedback

Methodology:

Trained for 3 epochs using qlora on Axolotl; the model uses ChatML prompt format.

Context Length:

16000

Training Time:

3 days

Hardware Used:

7x A6000s

Model Architecture:

Based on Experiment-26 by Yam Peleg

Input Output

Input Format:

ChatML prompt format

Accepted Modalities:

text

Output Format:

text

LLM Name	Dolphin 2.8 Experiment26 7B
Repository 🤗	https://huggingface.co/cognitivecomputations/dolphin-2.8-experiment26-7b
Model Size	7b
Required VRAM	14.4 GB
Updated	2025-02-22
Maintainer	cognitivecomputations
Model Type	mistral
Model Files	4.9 GB: 1-of-3 5.0 GB: 2-of-3 4.5 GB: 3-of-3
Supported Languages	en
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.39.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
Vocabulary Size	32002
Torch Data Type	float16

Quantized Models of the Dolphin 2.8 Experiment26 7B

Model	Likes	Downloads	VRAM
...periment26 7B Bnb 4bit Smashed	0	82	4 GB
...olphin 2.8 Experiment26 7B AWQ	1	89	4 GB

Best Alternatives to Dolphin 2.8 Experiment26 7B

Best Alternatives	Context / RAM	Downloads	Likes
...Nemo Instruct 2407 Abliterated	1000K / 24.5 GB	4620	11
MegaBeam Mistral 7B 512K	512K / 14.4 GB	5681	50
SpydazWeb AI HumanAI RP	512K / 14.4 GB	12	1
SpydazWeb AI HumanAI 002	512K / 14.4 GB	18	1
...daz Web AI ChatML 512K Project	512K / 14.5 GB	12	0
MegaBeam Mistral 7B 300K	282K / 14.4 GB	5633	16
Hebrew Mistral 7B 200K	256K / 30 GB	14619	15
Astral 256K 7B V2	250K / 14.4 GB	7	0
Astral 256K 7B	250K / 14.4 GB	5	0
Test001	128K / 14.5 GB	9	0

Note: green Score (e.g. "73.2") means that the model is better than cognitivecomputations/dolphin-2.8-experiment26-7b.

Rank the Dolphin 2.8 Experiment26 7B Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 43470 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Release v20241227

Support LLM Explorer