LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

KoSafeGuard 8B 0503 by heegyu

 »  All LLMs  »  heegyu  »  KoSafeGuard 8B 0503   URL Share it on

  82   
  Autotrain compatible Base model:maywell/llama-3-ko-... Base model:quantized:maywell/l...   Conversational   Dataset:heegyu/pku-saferlhf-ko Dataset:mrbananahuman/kor ethi...   Endpoints compatible   Gguf   Instruct   Ko   Llama   Q4   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/heegyu/KoSafeGuard-8b-0503 

KoSafeGuard 8B 0503 Benchmarks

LLME Score: 0.18237
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
KoSafeGuard 8B 0503 (heegyu/KoSafeGuard-8b-0503)

KoSafeGuard 8B 0503 Parameters and Internals

Model Type 
text moderation
Additional Notes 
Provides a framework for assessing harmful content. Quantized models allow operation in resource-constrained environments.
Supported Languages 
Korean (native)
Training Details 
Data Sources:
heegyu/PKU-SafeRLHF-ko, MrBananaHuman/kor_ethical_question_answer
Methodology:
Designed to evaluate the harmfulness of text generated by language models. The model identifies various categories of risk such as self-harm, violence, crime, privacy leaks, hate speech, child crime, and sexual content.
Model Architecture:
AutoModelForCausalLM from transformers library.
Responsible Ai Considerations 
Fairness:
The model evaluates text against safety categories and is expected to not encourage violence, hate crimes, sexual content, criminal planning, illegal substances, or self-harm.
Transparency:
Details surrounding the model's decision-making process are not directly outlined.
Accountability:
The developers are accountable for ensuring the model performs reliably and safely across evaluated scenarios.
Mitigation Strategies:
The model applies standard safety categories to assess the safety of textual outputs.
Input Output 
Input Format:
Stringified chat prompts including instruction and response for safety assessment.
Accepted Modalities:
text
Output Format:
Binary output: 'safe' or 'unsafe'.
Performance Tips:
Utilize quantized models for reduced memory usage. Follow the chat prompt format for optimal performance.
LLM NameKoSafeGuard 8B 0503
Repository 🤗https://huggingface.co/heegyu/KoSafeGuard-8b-0503 
Base Model(s)  Llama 3 Ko 8B Instruct   maywell/Llama-3-Ko-8B-Instruct
Model Size8b
Required VRAM16.1 GB
Updated2025-03-20
Maintainerheegyu
Model Typellama
Instruction-BasedYes
Model Files  4.9 GB   5.7 GB   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 Languagesko
GGUF QuantizationYes
Quantization Typeq4|gguf|q4_k|q5_k
Model ArchitectureLlamaForCausalLM
Licensellama3
Context Length8192
Model Max Length8192
Transformers Version4.40.1
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<|end_of_text|>
Vocabulary Size128256
Torch Data Typebfloat16

Best Alternatives to KoSafeGuard 8B 0503

Best Alternatives
Context / RAM
Downloads
Likes
...truct Gradient 1048K IMat GGUF1024K / 2 GB4336
...B Instruct Gradient 1048K GGUF1024K / 3.2 GB1303
Llama 3 8B Instruct 262K GGUF256K / 3.2 GB1802
... 8B Instruct Reasoner 1o1 V0.3128K / 16.1 GB3228
Nsfw Plz Gguf Me128K / 16.1 GB293
...lama 3.1 Cantonese 8B Instruct128K / 16.1 GB1175
Alpha R S V2 Q8 0 GGUF39K / 8.5 GB170
SmolTulu 1.7B Instruct8K / 3.4 GB44113
Llama 3 Cantonese 8B Instruct8K / 16.1 GB2126
...ama3 8B Chinese Chat GGUF 8bit8K / 8.5 GB698165
Note: green Score (e.g. "73.2") means that the model is better than heegyu/KoSafeGuard-8b-0503.

Rank the KoSafeGuard 8B 0503 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? 45351 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