Imo the biggest open research area in LLMs now is how to exactly do evaluation correctly. It's tricky as pointed out below
LLMS
-
Transformers: Six Years of Universal Neural Network Architecture
By
–
The "Attention is all you need paper" that introduced Transformer neural network architecture has been around for roughly six years. It's by far the first architecture to maintain its universality for a long time, not just for a single modality but for other modalities as well.
-
Build LLM Applications with Abacus AI Platform
By
–
REGISTER NOW FOR THIS https://
eventbrite.com/e/611420795057 Learn how, then build your own LLM applications using the @AbacusAI #AI-Assisted #DataScience platform—
Applications like:
• ChatLLM — search your own custom knowledge base
• DataLLM — search your tabular data & custom reports -
Building Games with LLM-Driven Non-Player Characters
By
–
We are actually building a game with LLM driven NPCs https://t.co/LDaBjAdTbt
— Marek Rosa | European🇪🇺 | South African🇿🇦 (@marek_rosa) 16 avril 2023We are actually building a game with LLM driven NPCs
-
Customizing Embeddings and Comparing Distance Metrics in AI
By
–
Nice — curious how this compares to learning a modified distance metric as shown here: https://
github.com/openai/openai-
cookbook/blob/main/examples/Customizing_embeddings.ipynb
… -
LLMO: The Next Wave of SEO for Large Language Models
By
–
This feels like the next wave of SEO. "LLMO"?(
https://
jina.ai/news/seo-is-de
ad-long-live-llmo/
…). It's as much about populating LLMs w/ good content as gaming them. If developers use LLMs to improve their work (and they are), how do you make sure they have the best possible path to using your code, well? -
Decoding latest AI news: GPT5, Chinese ChatGPT, offer and Twitter
By
–
We decode the latest #AI news → https://
youtu.be/on5PJq82zgg → Latest info on #GPT5 release
→ China launches its #ChatGPT
→ @OpenAI offers you $20,000
→ @ElonMusk launches #Twitter into AI! -
API temperature adjustment reduces AI hallucinations
By
–
c'est vrai que pour certains trucs il est assez déterministe par rapport a l'api, je pense qu'ils ont réglé la température plus bas qu'avant pour fixer au max l'hallucination
-
Red-teaming GPT-4: Insights into AI safety risks
By
–
Soon after GPT-4 came out I spoke to @paul_rottger for a story, who sparked the idea to speak to as many “red-teamers” as possible. Here’s the fruit of that reporting which hopefully gives some insight into the inner workings of this complex technology – and how it could go wrong