AI LLMs.
Collection of interesting LLM finetunes spanning across several topics and areas of expertise.
chain of thought

agents

SuperAGI/SAM Demo on Replicate 
Small Agentic Model that demonstrates impressive reasoning abilities despite its smaller size

Small Agentic Model that demonstrates impressive reasoning abilities despite its smaller size
WhiteRabbitNeo/Trinity-13B
Create autonomous agents
Create autonomous agents
Reasoning evaluation

allenai/digital-socrates-13b Demo on Replicate 
Digital Socrates is an open-source, automatic explanation-critiquing model

Digital Socrates is an open-source, automatic explanation-critiquing model
Overthinker
TheBloke/Sydney_Overthinker_13B
An over-analytical model
An over-analytical model
Input-output safeguard

llamas-community/LlamaGuard-7b Demo on Replicate 
Used for classifying content in both LLM inputs (prompt classification) and in LLM responses (response classification

Used for classifying content in both LLM inputs (prompt classification) and in LLM responses (response classification
model evaluation

kaist-ai/prometheus-7b-v1.0 Demo on Replicate 
An alternative to GPT-4 when evaluating LLMs & Reward models for RLHF

An alternative to GPT-4 when evaluating LLMs & Reward models for RLHF
dataset generation
NousResearch/Genstruct-7B
Create valid, synthetic instructions dataset for finetuning given a raw text corpus
Create valid, synthetic instructions dataset for finetuning given a raw text corpus
function calling

Nexusflow/NexusRaven-V2-13B Demo on Replicate 
Surpassing the state-of-the-art in open-source function calling LLMs

Surpassing the state-of-the-art in open-source function calling LLMs

gorilla-llm/gorilla-openfunctions-v1 Demo on Replicate 
Extend Chat Completion to formulate executable APIs call given natural language instructions and API context

Extend Chat Completion to formulate executable APIs call given natural language instructions and API context
meetkai/functionary-medium-v2.2
Interpret and execute functions/plugins
Interpret and execute functions/plugins
retrieval-augmented generation (RAG)

SciPhi/Sensei-7B-V1 Demo on Replicate 
Sensei is specialised in performing RAG over detailed web search results

Sensei is specialised in performing RAG over detailed web search results

Arc53/docsgpt-7b-mistral Demo on Replicate 
DocsGPT is optimized for Documentation (RAG), fine-tuned for providing answers that are based on context

DocsGPT is optimized for Documentation (RAG), fine-tuned for providing answers that are based on context
llmware/bling-phi-2-v0
Best Little Instruct No GPU Required
Best Little Instruct No GPU Required
llmware/dragon-mistral-7b-v0
Delivering RAG On Mistral
Delivering RAG On Mistral
benchmark while finetuning
Gryphe/MythoMist-7b
Actively benchmarks the model as it's being built
Actively benchmarks the model as it's being built
megamix

KoboldAI/LLaMA2-13B-Tiefighter Demo on Replicate 
A merged model achieved trough merging two different lora's on top of a well established existing merge

A merged model achieved trough merging two different lora's on top of a well established existing merge
CalderaAI/Naberius-7B
Uncensored, Pliant, Logic-Based, & Imaginative Instruct-Based Spherically Interpolated Tri-Merge
Uncensored, Pliant, Logic-Based, & Imaginative Instruct-Based Spherically Interpolated Tri-Merge
EmbeddedLLM/Mistral-7B-Merge-14-v0.1
This is an experiment to test merging 14 models using DARE TIES 🦙
This is an experiment to test merging 14 models using DARE TIES 🦙
Frankenmerges
athirdpath/BigLlama-20b-v1.1
Merge 4 Llama-13b into a 20b model
Merge 4 Llama-13b into a 20b model
Sao10K/Solus-103B-L2
Experimental 100B Versions. Better than 70b models, without the spelling/number issues 120b models like Goliath had
Experimental 100B Versions. Better than 70b models, without the spelling/number issues 120b models like Goliath had
alpindale/goliath-120b
An auto-regressive causal LM created by combining 2x finetuned Llama-2 70B into one
An auto-regressive causal LM created by combining 2x finetuned Llama-2 70B into one
jan-ai/Pandora-13B-v1
This model uses the passthrough merge method from the best 7B models
This model uses the passthrough merge method from the best 7B models
Llama trained on Claude 2 chats

umd-zhou-lab/claude2-alpaca-13B Demo on Replicate 
This model is trained by fine-tuning llama-2 with claude2 alpaca data

This model is trained by fine-tuning llama-2 with claude2 alpaca data
Time Travel?

Pclanglais/MonadGPT Demo on Replicate 
What would have happened if ChatGPT was invented in the 17th century?

What would have happened if ChatGPT was invented in the 17th century?
esoteric, occult, and spiritual

teknium/Mistral-Trismegistus-7B Demo on Replicate 
Mistral Trismegistus is a model made for people interested in the esoteric, occult, and spiritual

Mistral Trismegistus is a model made for people interested in the esoteric, occult, and spiritual
evil tuned models
maywell/PiVoT-0.1-Evil-a
Reckless, Evil Assistant
Reckless, Evil Assistant
Gryphe/Tiamat-7b
A five-headed dragon goddess embodying wickedness and cruelty from the Forgotten Realms
A five-headed dragon goddess embodying wickedness and cruelty from the Forgotten Realms
Undi95/toxicqa-Llama2-7B
This model is based on a toxic dataset
This model is based on a toxic dataset
text based adventure

PocketDoc/Dans-AdventurousWinds-Mk2-7b Demo on Replicate 
This model is proficient in crafting text-based adventure games

This model is proficient in crafting text-based adventure games
Role-playing
Sao10K/Ana-v1-m7
A model solely focused on the RP / ERP Experience.
A model solely focused on the RP / ERP Experience.
KoboldAI/LLaMA2-13B-Estopia
Focused on "guided narratives"
Focused on "guided narratives"
Delcos/Velara-11B-V2
A model focused on being an assistant worth talking to
A model focused on being an assistant worth talking to
Norquinal/OpenCAI-7B
Open-source recreation of the style of roleplay found at C.AI
Open-source recreation of the style of roleplay found at C.AI
mathematics
akjindal53244/Arithmo-Mistral-7B
Trained to reason and answer mathematical problems
Trained to reason and answer mathematical problems
EleutherAI/llemma_34b
Particularly strong at chain-of-thought mathematical reasoning
Particularly strong at chain-of-thought mathematical reasoning

meta-math/MetaMath-Mistral-7B Demo on Replicate 
Bootstrap Your Own Mathematical Questions for Large Language Models

Bootstrap Your Own Mathematical Questions for Large Language Models
data analysis

pipizhao/Pandalyst-7B-V1.2 Demo on Replicate 
Pandalyst is a large language model for mastering data analysis using pandas

Pandalyst is a large language model for mastering data analysis using pandas
medicine
AdaptLLM/medicine-chat
Finetuned on medicine knowledge
Finetuned on medicine knowledge
Severus27/BeingWell_llama2_7b
Trained on a dataset comprising USMLE (United States Medical Licensing Examination) questions and answers
Trained on a dataset comprising USMLE (United States Medical Licensing Examination) questions and answers
sethuiyer/Dr_Samantha-7b
Has capabilities of a medical knowledge-focused model with the philosophical, psychological, and relational understanding of the Samantha-7b model
Has capabilities of a medical knowledge-focused model with the philosophical, psychological, and relational understanding of the Samantha-7b model
BioMistral/BioMistral-7B
Suited for medical domains pre-trained using textual data from PubMed Central Open Access
Suited for medical domains pre-trained using textual data from PubMed Central Open Access
mental health

steve-cse/MelloGPT
A large language model fine-tuned on mental health counseling conversations
A large language model fine-tuned on mental health counseling conversations
electrical engineering
STEM-AI-mtl/phi-2-electrical-engineering
Q&A related to electrical engineering, and Kicad software. Creation of Python code in general, and for Kicad's scripting console
Q&A related to electrical engineering, and Kicad software. Creation of Python code in general, and for Kicad's scripting console
cybersecurity

WhiteRabbitNeo/WhiteRabbitNeo-13B-v1 Demo on Replicate 
WhiteRabbitNeo is a model series that can be used for offensive and defensive cybersecurity

WhiteRabbitNeo is a model series that can be used for offensive and defensive cybersecurity
programming
defog/sqlcoder-7b-2
Natural language to SQL generation
Natural language to SQL generation
translation and multilingual

haoranxu/ALMA-13B Demo on Replicate 
ALMA (Advanced Language Model-based trAnslator) is an LLM-based translation model

ALMA (Advanced Language Model-based trAnslator) is an LLM-based translation model


Unbabel/TowerInstruct-7B-v0.1 Demo on Replicate 
This model is trained to handle several translation-related tasks, such as general machine translation, gramatical error correction, and paraphrase generation

This model is trained to handle several translation-related tasks, such as general machine translation, gramatical error correction, and paraphrase generation
language specific models
projecte-aina/FLOR-1.3B-Instructed
Catalan, Spanish, and English
Catalan, Spanish, and English
Rijgersberg/GEITje-7B-chat
Dutch language skills and knowledge of Dutch topics
Dutch language skills and knowledge of Dutch topics
LumiOpen/Poro-34B
Finnish, English and code
Finnish, English and code
croissantllm/CroissantLLMChat-v0.1
French and English
French and English
Telugu-LLM-Labs/Indic-gemma-2b-finetuned-sft-Navarasa
9 Indian languages (Hindi, Telugu, Tamil, Kannada, Malayalam, Gujarati, Punjabi, Bengali, Odia) and English
9 Indian languages (Hindi, Telugu, Tamil, Kannada, Malayalam, Gujarati, Punjabi, Bengali, Odia) and English
maywell/koOpenChat-sft
Korean
Korean
NorGLM/NorLlama-3B
Norwegian, Denish, Swedish, Germany and English
Norwegian, Denish, Swedish, Germany and English
eryk-mazus/polka-1.1b-chat
Polish
Polish
lrds-code/samba-1.1B
Portuguese (Angola, Brazil, Cape Verde, Guinea-Bissau, Equatorial Guinea, Mozambique, Portugal, São Tomé and Príncipe, Timor-Leste)
Portuguese (Angola, Brazil, Cape Verde, Guinea-Bissau, Equatorial Guinea, Mozambique, Portugal, São Tomé and Príncipe, Timor-Leste)
IlyaGusev/saiga_gemma_9b
Russian
Russian
SeaLLMs/SeaLLM-7B-v2
South-asian (Vietnamese, Indonesian, Thai, Malay, Khmer, Lao, Tagalog and Burmese)
South-asian (Vietnamese, Indonesian, Thai, Malay, Khmer, Lao, Tagalog and Burmese)
AI-Sweden-Models/gpt-sw3-1.3b-instruct
Swedish, Norwegian, Danish, Icelandic, English
Swedish, Norwegian, Danish, Icelandic, English
scb10x/typhoon-7b
Thai
Thai
asafaya/kanarya-2b
Turkish
Turkish
vilm/vinallama-7b-chat
Vietnamese
Vietnamese
SLM (small language models)
NucleusOrg/Nucleus-1B-alpha-1
Small language model based on Mistral (trimmed untrained version)
Small language model based on Mistral (trimmed untrained version)
TinyLlama/TinyLlama-1.1B-Chat-v1.0
1.1B Llama model on 3 trillion tokens
1.1B Llama model on 3 trillion tokens
HuggingFaceTB/cosmo-1b
1.8B model trained on Cosmopedia synthetic dataset
1.8B model trained on Cosmopedia synthetic dataset
multimodal llm

NousResearch/Obsidian-3B-V0.5 Demo on Replicate 
Worlds smallest multi-modal LLM (open source gpt4 vision)

Worlds smallest multi-modal LLM (open source gpt4 vision)