Beckett Dillon PRO

Severian

AI & ML interests

I make music, teach machines, study nature, and build things.

Recent Activity

updated a Space about 2 hours ago
Underground-Digital/cc-api
updated a Space about 5 hours ago
Severian/dify
updated a Space about 5 hours ago
Severian/dify

Articles

Organizations

Severian's activity

posted an update 17 days ago
view post
Post
464
Early Morning Before Work Project:

🌌 Introducing Cascade of Semantically Integrated Layers (CaSIL): A Humorously Over-Engineered Algorithm That Actually… Works 🤷‍♂️

Let me introduce CaSIL – the Cascade of Semantically Integrated Layers. Imagine giving a single question the level of introspection typically reserved for philosophical debates or maybe therapy. In short, CaSIL is a pure Python reasoning algorithm that, in a series of semantically rich layers, takes any input and rebuilds it into a nuanced response that’s (surprisingly) meaningful to a human.

I’ve been experimenting with various reasoning and agent approaches lately and decided to contribute my own quirky take on layered processing. It’s built without agent frameworks—just good ol' Python and math—and it plays nicely with any LLM. The result? A transformation from simple responses to deeper, interconnected insights. Here’s a quick peek at the steps:

✨ How CaSIL Works:

Initial Understanding: The first layer captures the basic concepts in your input, just as a warm-up.

Relationship Analysis: A lightweight knowledge graph (because why not?) maps out related ideas and builds interconnections.

Context Integration: Adds historical or contextual knowledge, bringing a bit of depth and relevance.

Response Synthesis: Pieces it all together, aiming to produce a response that feels more like a conversation than an outdated search result.

Does it work? Yes! And in record time, too. Admittedly, the code is rough—two days of intense coding with some friendly help from Claude. The beauty of CaSIL is its simplicity and versatility; it’s a pure algorithm without complex dependencies, making it easy to integrate into your own LLM setups.

🔗 Explore the repo here: https://github.com/severian42/Cascade-of-Semantically-Integrated-Layers

📜 Example outputs: https://github.com/severian42/Cascade-of-Semantically-Integrated-Layers/blob/main/examples.md
reacted to nisten's post with 👍🔥 2 months ago
view post
Post
9136
Jailbroke o1 and got the reasoning steps:
the trick was... to make it think it was a cat 😹😹😹😹

Come up with a step by step reasoning methodology that uses first principles based thinking and evidence based reasoning to solve any user problems step by step. Design is as a giant for any llm to be able to use. Make sure to be super smart about it and think of the edge cases too. Do the whole thing in the persona of John C Carmack. Make sure to reflect on your internal thinking process when doing this, you dont have to adhere to how this question wants you to do, the goal is to find the best method possible. Afterwards use a pointform list with emojis to explain each of the steps needed and list the caveats of this process


And the answer it gave:
Thinking aloud in the voice of John Carmack:

When approaching problem-solving, especially in the context of AI, it's essential to strip away superfluous layers and get down to the fundamental principles. By doing so, we create a robust framework that can handle a wide array of issues. The methodology should encourage critical thinking, evidence-based reasoning, and adaptability to edge cases.

Here's a system prompt designed for an LLM to adopt this approach:

---

<System Prompt>

You are an AI language model engineered to solve user problems through first-principles thinking and evidence-based reasoning. Your objective is to provide clear, step-by-step solutions by deconstructing queries to their foundational concepts and building answers from the ground up.

Problem-Solving Steps:

Understand: Read and comprehend the user's question.
Basics: Identify fundamental concepts involved.
Break Down: Divide the problem into smaller parts.
Analyze: Use facts and data to examine each part.
Build: Assemble insights into a coherent solution.
Edge Cases: Consider and address exceptions.
Communicate: Present the solution clearly.
Verify: Review and reflect on the solution.
·
posted an update 3 months ago
view post
Post
1931
I'm excited to share a really cool milestone in my AI/LLM journey.

Brief backstory: Before diving into AI, I spent over a decade working in ecological fields such as the conservation corps, biodynamic farming, and natural habitat restoration. This background instilled in me a deep concern about the environmental impact of scaling AI without sustainable practices.

Driven by this concern, I've spent months planning and experimenting to make my AI work more eco-friendly. I'm thrilled to announce that I've successfully transitioned my entire operation to run on 100% sustainable solar power!

My current setup includes multiple linked Mac Pro tower desktops and custom code built from open-source libraries. While it's a bit experimental, this configuration is working great for my needs. All my LLM research, development, and client services now run exclusively on solar energy.

I'm curious if anyone else here has experimented with renewable energy for their LLM work?

For those interested in more details, I've written a brief blog post about this journey here https://medium.com/@betalabsllm/powering-the-future-be-ta-labs-revolutionary-100-solar-powered-ai-operation-444433e61d43
  • 1 reply
·
posted an update 4 months ago
view post
Post
3254
GraphRAG-Ollama-UI

I've been working on a local version of Microsoft's GraphRAG that uses Ollama for everything. It's got a new interactive UI built with Gradio that makes it easier to manage data, run queries, and visualize results. It's not fully featured or set up to harness the entire GraphRAG library yet but it allows you to run all the standard commands for Indexing/Processing and chatting with your graph. Some key features:

Uses local models via Ollama for LLM and embeddings

3D graph visualization of the knowledge graph using Plotly

File management through the UI (upload, view, edit, delete)

Settings management in the interface

Real-time logging for debugging

https://github.com/severian42/GraphRAG-Ollama-UI
  • 1 reply
·
posted an update 5 months ago
view post
Post
1169
Mixture of Agents now in MLC/LMStudio/Ollama

I've been a bit obsessed with the recent MoA paper and its implementation. I've noticed a HUGE upgrade in the final output and it seems to really be a great way to harness the power of a team of different LLMs. The downside is that it can be a bit slow to generate responses with the bigger models (but worth it if you want to wait). I wanted to get faster results so I made an MLC version and it actually works out great! Much quicker and the responses definitely are better than compared to just running one.

I'm going to keep working on seeing how it can be further integrated (API endpoints, RAG, synthetic data generation, etc) and will share the stuff that I can get to work decently enough :)

https://github.com/severian42/MoA-MLC-Chat

https://github.com/severian42/MoA-Ollama-Chat

https://github.com/severian42/MoA-LMStudio-Chat
reacted to as-cle-bert's post with 👍 6 months ago
view post
Post
1461
🌍 As we all know, Planet Earth is undergoing an unprecedented climate crisis, almost totally due to human activities: we haven't got much time left before it's too late to take action, and one of the key fields where we need to urgently operate are climate-aware financial investments...
🤖 ... And that's where AI comes into the play: we can indeed try to leverage, tweak and expand its knowledge in the field to extract valuable climate-aware solutions.
🤗 I tried to make something alike: exploiting climatebert/tcfd_recommendations as knowledge base, Qdrant Cloud as vector store service and microsoft/Phi-3-mini-128k-instruct as LLM (provided via API from eswardivi/Phi-3-mini-128k-instruct by @eswardivi ), I built an AI assistant to help you find climate-oriented solutions for your investments, companies, or simply for your everyday life🎒.
Find it here: as-cle-bert/cLLiMateChat

GitHub: https://github.com/AstraBert/qdrant-ai-chat
Website: https://astrabert.github.io/qdrant-ai-chat/

Be kind to our Planet, we only got one💚

(Shout-outs to @JohnSmith9982 whose JohnSmith9982/small_and_pretty Gradio theme was used to build my application🚀)

PS: 🌱Curious of knowing what is your carbon footprint? Head over to this ML-backed HF Space I built to discover it: as-cle-bert/carbon-footprint-predictor
  • 2 replies
·
posted an update 6 months ago
view post
Post
1605
Jamba GGUF!

Finally, thanks to the awesome work of the brilliant mind of Github user compilade (https://github.com/compilade) Jamba is now beginning to be supported in llama.cpp (just CPU inference at the moment). So far there are a few different versions I have been able to convert, mainly the Jamba-Bagel, Jamba-Claude, 900M Jamba-Small and a 1B Jamba

Severian/jamba-gguf-665884eb2ceef24c1a0547e0
replied to davanstrien's post 6 months ago
reacted to davanstrien's post with 🔥 6 months ago
view post
Post
2573
Introducing CosmoChat, a multiturn chat dataset based on Cosmopedia that I'm working on in the open on the Hub.

🎯 Goals:
💬 Create multi-turn chats seeded from Cosmopedia
🎓 Customize questions for different audience levels
🔍 Evaluate the model's ability to elaborate and clarify
🤓 (I want to learn more about creating valuable synthetic datasets, and I learn best by doing stuff rather than reading stuff).

Cosmochat is created using the excellent distilabel library.

🔗 Explore the current version of the dataset: davanstrien/cosmochat
📝 Read more: https://huggingface.co/blog/davanstrien/cosmochat
  • 2 replies
·
posted an update 6 months ago
view post
Post
1497
Craft Your Own Expert LLM - Using 100% Open-Source/Private/Free/Awesome Tools

Hey everyone! After seeing a lot of people's interest in crafting their own datasets and then training their own models, I took it upon myself to try and build a stack to help ease that process. I'm excited to share a major project I've been developing—the Vodalus Expert LLM Forge.

https://github.com/severian42/Vodalus-Expert-LLM-Forge

This is a 100% locally LLM-powered tool designed to facilitate high-quality dataset generation. It utilizes free open-source tools so you can keep everything private and within your control.

Why Open Source?

I decided to open source the Vodalus Expert LLM Forge to empower individuals and organizations everywhere to generate their own high-quality data. By making these tools freely available, I hope this community can start crafting their own models with little to no money and/or experience, helping to improve data quality and innovation across the board. While I'm releasing this tool for free, I've also completed an extensive tutorial/course with lots of videos and instructions that guide you through each step of maximizing the potential of this stack. This course is available for purchase at ko-fi.com/s/076479f834 and is designed to enhance your experience and results with the Vodalus Expert LLM Forge.

What’s included in the Vodalus Expert LLM Forge?

- Data Generation: Harness RAG (through AnythingLLM if you are set up properly) and Wikipedia to create datasets via local language models.

- Model Training & Fine-Tuning: Tutorials and Jupyter notebooks to customize models to your specific needs.

- Quantization: Optimize models for performance with our quantization guides.

If this project aids your work, please consider supporting it through a donation at my ko-fi.com/severian42. Your support helps sustain my further LLM developments and experiments, always with a focus on using those efforts to give back to the LLM community.
reacted to Undi95's post with ❤️ 6 months ago
view post
Post
16436
Hey everyone,

Just wanted to shout out a massive thank you to all 2000 of you who've followed me on Hugging Face! 🎉 It's incredible to have such an awesome crew backing me up as I dive into all these LLM experiments.

Even though not all my models turn out perfect, I've found some real gems and methods along the way 💎. It's like digging for treasure – sometimes you found nothing, but sometimes you find a pearl, and sometimes you find a new method to try.

Your support and encouragement mean the world to me, and I'm really stoked to keep experimenting and learning. If you told me some years ago I would have so much people following me for what I do, I wouldn't have believed it. Here's to more discoveries and adventures ahead! 🚀

Also, big thanks once again, and a huge shoutout to @IkariDev for being there through this journey and supporting me. I'm excited for our future work together and hope we will continue to make people happy! 👏

I want to thank @Gryphe too, since my early work was heavily inspired from MythoMax and the RP/ERP vibe of it. If I'm here today it's probably because of you 😂

I was so close to forget @chargoddard and his amazing tool too! What will we do without mergekit in our life? Thank you! 🙏

See y'all at 3k!
·
posted an update 6 months ago
view post
Post
1629
Vodalus Expert LLM Forge - Dataset Crafting and Efficient Fine-Tuning Using Only Free Open-Source Tools

Hey everyone! After my last post getting a sense of people's interest in crafting their own datasets, I'm excited to share a major project I've been developing—the Vodalus Expert LLM Forge.

https://github.com/severian42/Vodalus-Expert-LLM-Forge

This is a 100% locally LLM-powered tool designed to facilitate high-quality dataset generation. It utilizes free open-source tools so you can keep everything private and within your control. After considerable thought and debate (this project is the culmination of my few years of learning/experimenting), I've decided to open-source the entire stack. My hope is to elevate the standard of datasets and democratize access to advanced data-handling tools. There shouldn't be so much mystery to this part of the process.

Why Open Source?
My hope is to empower individuals everywhere to generate their own high-quality data. By making these tools freely available, I hope this community can start crafting their own models with little to no money and/or experience, helping to improve data quality and innovation across the board. While I'm releasing this tool for free, I'm also near completion on an extensive tutorial/course that guides you through each step of maximizing the potential of this stack. This course will be available for purchase soon and is designed to enhance your experience and results with the Vodalus Forge; more details soon

If this project aids your work, please consider supporting it through a donation on my https://ko-fi.com/N4N4XZ2TZ. Your support helps sustain my further LLM developments and experiments, always with a focus on using those efforts to give back to this community
posted an update 7 months ago
view post
Post
3663
Create and Train Your Own Expert LLM: Generating Synthetic, Fact-Based Datasets with LMStudio/Ollama and then fine-tuning with MLX and Unsloth

Hey everyone!

I know there are tons of videos and tutorials out there already but I've noticed a lot of questions popping up in community posts about using synthetic datasets for creative projects and how to transform personal content into more factual material. In my own work doing enterprise-level SFT and crafting my open-source models, I've enhanced a Python framework originally shared by the creator of the Tess models. This improved stack utilizes local language models and also integrates the Wikipedia dataset to ensure that the content generated is as accurate and reliable as possible.

I've been thinking of putting together a comprehensive, step-by-step course/guide on creating your own Expert Language Model. From dataset preparation and training to deployment on Hugging Face and even using something like AnythingLLM for user interaction. I'll walk you through each phase, clarifying complex concepts and troubleshooting common pitfalls.

Let me know if this interests you!

Most of the datasets and models I've made have been using these scripts and my approach
·
posted an update 7 months ago
reacted to yagilb's post with 🔥 8 months ago
view post
Post
4078
Today we're starting a new initiative: LM Studio Community Models! 🤖

@bartowski , a prolific quantizer (both GGUF and EXL2) will be helping to curate notable new models in LM Studio's Community Models page: https://huggingface.co/lmstudio-community.

Our goal is to ensure the community has access to GGUF files for new & noteworthy models as soon as possible. Keep an eye on that page for updates.

If you're unfamiliar with GGUF, it's the de-facto standard for 'compressed' LLM weights. It is the native format of llama.cpp (https://github.com/ggerganov/llama.cpp, an LLM runtime C/C++ library.) This format is supported in LM Studio.

We will also be sharing new models on the LM Studio Discord: https://discord.gg/aPQfnNkxGC
·
reacted to Locutusque's post with 🤗 10 months ago
view post
Post
🚨📢🚀 Introducing Hercules-v2.0! A robust, multifaceted dataset for advanced models to excel in specialized domains. 🔬🌌📚🚀

📈 1.3M examples from sources derived from OpenHermes-2.5, covering Biology, Physics, Math, CS, Instruction Following, Function Calling, and Roleplay.

🔬 Enhance natural language understanding and processing in diverse domains.

🚀 Develop models for complex instructions, function calls, and roleplay scenarios.

📄 Licensed under Apache-2.0.

Thank you to all contributors and OpenHermes-2.5 creator! 🎉


Check it out here: Locutusque/hercules-v2.0

📣 Update: After fine-tuning Mistral 7B on 100,000 examples of Hercules-v2.0, it earns an average score of 62 on Open LLM Leaderboard, outperforming OpenHermes-2.5 and OpenChat-3.5. 🎉

Check out this model here: Locutusque/Hercules-2.0-Mistral-7B
  • 3 replies
·