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

Luxia 21.4B Alignment V1.2 by saltlux

 »  All LLMs  »  saltlux  »  Luxia 21.4B Alignment V1.2   URL Share it on

  5553   
  Autotrain compatible   En   Endpoints compatible   Llama   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/saltlux/luxia-21.4b-alignment-v1.2 

Luxia 21.4B Alignment V1.2 Benchmarks

MMLU Pro: 27.48
GPQA: 7.72
MUSR: 14.9
BBH: 47.77
IFEval: 41.15 vs 88 (so35)-53.2%
ARC: 77.73 vs 96.7 (so35)-19.6%
HellaSwag: 90.86 vs 95.3 (gpt4)-4.7%
MMLU: 67.86 vs 88.3 (so35)-23.1%
TruthfulQA: 79.16 vs 59 (gpt4)34.2%
WinoGrande: 86.27 vs 87.5 (gpt4)-1.4%
GSM8K: 66.94 vs 96.4 (so35)-30.6%
MATH Lvl 5: 8.46
LLME Score: 0.30358
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Luxia 21.4B Alignment V1.2 (saltlux/luxia-21.4b-alignment-v1.2)

Luxia 21.4B Alignment V1.2 Parameters and Internals

Model Type 
large language model
Additional Notes 
Model demonstrates state-of-the-art performance in models with parameters under 35B
Supported Languages 
en (proficient)
Training Details 
Data Sources:
c-s-ale/alpaca-gpt4-data, Open-Orca/SlimOrca, jondurbin/truthy-dpo-v0.1, abacusai/ARC_DPO_FewShot, abacusai/HellaSwag_DPO_FewShot
Methodology:
Supervised fine-tuning (SFT), Direct Preference Optimization (DPO)
Model Architecture:
Expanded layers through passthrough, continual pretraining based on internlm2-20b-llama
LLM NameLuxia 21.4B Alignment V1.2
Repository 🤗https://huggingface.co/saltlux/luxia-21.4b-alignment-v1.2 
Model Size21.4b
Required VRAM43 GB
Updated2025-02-22
Maintainersaltlux
Model Typellama
Model Files  4.8 GB: 1-of-9   5.0 GB: 2-of-9   5.0 GB: 3-of-9   4.9 GB: 4-of-9   5.0 GB: 5-of-9   5.0 GB: 6-of-9   4.9 GB: 7-of-9   5.0 GB: 8-of-9   3.4 GB: 9-of-9
Supported Languagesen
Model ArchitectureLlamaForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.37.2
Is Biased0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size92544
Torch Data Typebfloat16

Best Alternatives to Luxia 21.4B Alignment V1.2

Best Alternatives
Context / RAM
Downloads
Likes
Luxia 21.4B Alignment V1.032K / 42.8 GB968633
UNA ThePitbull 21.4B V232K / 43 GB4316
Final Model Test V232K / 43 GB100
DPO Model Test132K / 43 GB50
UNA ThePitbull 21.4 V132K / 43 GB195
Sft Model Test132K / 43 GB100
Merge Model Test232K / 42.8 GB80
Merge Model Test132K / 42.8 GB60
Sft Model Test032K / 43 GB70
All DPO Model Test132K / 43 GB70

Rank the Luxia 21.4B Alignment V1.2 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