almost 30k followers on HF https://
huggingface.co/deepseek-ai
LLMS
-
DeepSeek Reaches 30k Followers on Hugging Face
By
–
-
Expert LLMs Don’t Require Reasoning Capabilities
By
–
The one thing to add though is that an "expert" LLM doesn't necessarily have to be a reasoning LLM (you could have one that answers biology Q&As perfectly without reasoning steps involved).
Similarly, you can have a generalist reasoning LLM that is not an expert in one -
Rule-Based Rewards Scale Better Than Style Preferences
By
–
The nice thing here is that the rule-based rewards scale better. And for things like code and math, they also make a lot more sense. I.e., you care more about correctness than style preference. Btw @natolambert 's team's OLMo 2 also used verifiable rewards in the RLHF stage for
-
From Prompt Engineering to Personality Engineering: 2025 Trends
By
–
2024: all about prompt engineering
2025: all about personality engineering -
Google Launches Gemini 2.0 Pro: Strongest Model with 2M Token Context
By
–
3/ Google’s Gemini 2.0 Pro arrived—their strongest model yet It’s Google's best model for coding performance and complex prompts, with a 2M token context window, the longest in its category. Available now in Google AI Studio, Vertex AI, and Gemini App Advanced.
-
OpenAI Launches o3-mini: Cost-Efficient Reasoning Model for Science and Math
By
–
4/ OpenAI launches o3-mini, its most cost-efficient reasoning model yet o3-mini is optimized for science, math, and coding with fast speed and precision. Many in the community find it impressive, even better than o1 in some use cases. o3-mini is available in ChatGPT.
-
Mistral Small 3: Fast Open-Source Model Competes with Larger Models
By
–
5/ Mistral Small 3: fast, open-source, and ready to compete Mistral released Small 3, a 24B model with 81% MMLU accuracy, impressing the community with its performance and efficiency in the small model category. It can even compete with models 3x its size.
-
Top AI News from the Week: OpenAI, GitHub Copilot, Mistral, Google
By
–
Top AI news from this week that you can't miss, We summarized everything from OpenAI, GitHub Copilot, Mistral, Google, and more. Here's everything you need to know:
-
Mistral Launches Le Chat with Advanced Features and Speed
By
–
2/ Mistral's new Le Chat has been released:
— AlphaSignal AI (@AlphaSignalAI) 7 février 2025
– Blazing speed: Processes up to 1,100 tokens/sec.
– New tools: Canvas, web search, image generation, code interpreter.
– Advanced OCR: Analyzes PDFs, spreadsheets, and scanned docs.
Available for free on iOS, Android, and web. pic.twitter.com/LESJluesFX2/ Mistral's new Le Chat has been released: – Blazing speed: Processes up to 1,100 tokens/sec.
– New tools: Canvas, web search, image generation, code interpreter.
– Advanced OCR: Analyzes PDFs, spreadsheets, and scanned docs. Available for free on iOS, Android, and web. -
GraphRAG Outperforms Standard RAG with Entity Relationships
By
–
GraphRAG outperforms Standard RAG. It improves retrieval by using relationships between entities, leading to more accurate and context-aware responses. This makes it more effective than standard RAG. At Abacus, we automatically build the best RAG system for big data.