Meta Llama 3 8B Instruct By meta-llama: Benchmarks, Features and Detailed Analysis. Insights on Meta Llama 3 8B Instruct.

Autotrain compatible Conversational En Endpoints compatible Facebook Instruct Llama Llama-3 Meta Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct

Meta Llama 3 8B Instruct Benchmarks

MMLU Pro: 28.79

GPQA: 5.7

MUSR: 5.4

BBH: 26.8

IFEval: 47.82 vs 88 (so35)^-45.7%

ARC: 60.75 vs 96.7 (so35)^-37.2%

HellaSwag: 78.55 vs 95.3 (gpt4)^-17.6%

MMLU: 67.07 vs 88.3 (so35)^-24%

TruthfulQA: 51.65 vs 59 (gpt4)^-12.5%

WinoGrande: 74.51 vs 87.5 (gpt4)^-14.8%

GSM8K: 68.69 vs 96.4 (so35)^-28.7%

MATH Lvl 5: 9.14

LLME Score: 0.47452

^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)

Meta Llama 3 8B Instruct (meta-llama/Meta-Llama-3-8B-Instruct)

Meta Llama 3 8B Instruct Parameters and Internals

Model Type

text generation, dialogue optimization

Use Cases

Areas:

research, commercial applications

Applications:

assistant-like chat, natural language generation

Primary Use Cases:

Instruction tuned for chat-oriented tasks, Pretrained model adaptation

Limitations:

English language focus

Considerations:

Follow Responsible Use Guide and implement safety tools.

Additional Notes

Developed with openness, inclusivity, and helpfulness in mind.

Supported Languages

English (Fluent)

Training Details

Data Sources:

publicly available online data

Data Volume:

15 trillion tokens

Methodology:

Supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF)

Context Length:

8000

Training Time:

7.7 million GPU hours

Hardware Used:

H100-80GB GPUs

Model Architecture:

Auto-regressive transformer

Safety Evaluation

Methodologies:

red-teaming, adversarial evaluations

Findings:

Improved safety measures than previous versions

Risk Categories:

misinformation, cybersecurity, child safety

Ethical Considerations:

Ensures transparency and responsibility in deployment.

Responsible Ai Considerations

Fairness:

Developers are encouraged to implement safety tools and assess risks.

Transparency:

Open source tooling and contribution encouraged.

Accountability:

Meta provides safety benchmarks and regular updates.

Mitigation Strategies:

Implement safeguards using Meta Llama Guard.

Input Output

Input Format:

Text input

Accepted Modalities:

text

Output Format:

Generates text and code

Performance Tips:

Use specified transformers libraries for enhanced performance.

Release Notes

Version:

3.1

Date:

April 18, 2024

Notes:

Introduced new variants with fine-tuning optimizations.

LLM Name	Meta Llama 3 8B Instruct
Repository 🤗	https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct
Model Size	8b
Required VRAM	16.1 GB
Updated	2025-04-28
Maintainer	meta-llama
Model Type	llama
Instruction-Based	Yes
Model Files	5.0 GB: 1-of-4 5.0 GB: 2-of-4 4.9 GB: 3-of-4 1.2 GB: 4-of-4
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	llama3
Context Length	8192
Model Max Length	8192
Transformers Version	4.40.0.dev0
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	128256
Torch Data Type	bfloat16

Quantized Models of the Meta Llama 3 8B Instruct

Model	Likes	Downloads	VRAM
Meta Llama 3 8B Instruct GGUF	88	282702	2 GB
OrcaAgent Llama3.2 8B	4	9	16 GB
Llama 3 8B Instruct GPTQ 8 Bit	25	21167	9 GB
Ai Medical Model 32bit	1	2672	42 GB
Nimue 8B Merged	1	0	5 GB
Llama 3 8B Instruct GPTQ 4 Bit	25	2073	5 GB
...ama3 8B Chinese Chat GGUF 8bit	165	226	8 GB
Meta Llama 3 8B Instruct GPTQ	6	1778	5 GB
Meta Llama 3 8B Instruct AWQ	0	36	5 GB
...eta Llama 3 8B Instruct Zh 10K	0	14	16 GB

Best Alternatives to Meta Llama 3 8B Instruct

Best Alternatives	Context / RAM	Downloads	Likes
...otron 8B UltraLong 4M Instruct	4192K / 32.1 GB	4241	103
UltraLong Thinking	4192K / 16.1 GB	69	2
...a 3.1 8B UltraLong 4M Instruct	4192K / 32.1 GB	176	24
...otron 8B UltraLong 2M Instruct	2096K / 32.1 GB	1289	15
...a 3.1 8B UltraLong 2M Instruct	2096K / 32.1 GB	875	9
...otron 8B UltraLong 1M Instruct	1048K / 32.1 GB	3625	40
...a 3.1 8B UltraLong 1M Instruct	1048K / 32.1 GB	1387	29
....1 1million Ctx Dark Planet 8B	1048K / 32.3 GB	26	2
...a 3 8B Instruct Gradient 1048K	1024K / 16.1 GB	26621	679
A4.2	1024K / 16.1 GB	135	0

Rank the Meta Llama 3 8B Instruct 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? 46792 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