Update README.md

README.md

@@ -55,30 +55,16 @@ Processing single interactions in real-time by **Reactive Language Models** lead
 > will be about **15x cheaper**
 
 > Reactive Transformer architecture was analysed by 10 state-of-the-art LLM/Reasoning models for its innovations and market disruption potential,
-> rated as ~4.36/5.0. Check - [Reactive Transformer AI Analysis](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/ReactiveTransformer/ai-analysis.md)
-
-## Reactive Transformer - drafts
-- [Architecture introduction](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/ReactiveTransformer/reactive-transformer.md)
-- [Supervised Training stages](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/ReactiveTransformer/supervised-training.md)
-- [Reinforcement Learning stages](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/ReactiveTransformer/mrl.md)
-
-### RxT-Alpha Open Research
-We are currently working on **Reactive Transformer Proof-of-Concept - RxT-Alpha**, especially on the new reinforcement learning stage - **Memory Reinforcement Learning**,
-that's required for our reactive models, between the _Supervised Memory System Training (SMST)_ and _Reinforcement Learning from Human Feedback for reactive models (RxRLHF)_.
-The research is open, we are publishing the results of all separate steps, just after finishing them.
-
-We are currently finishing **MRL** training for the world's first experimental (Proof-of-Concept) reactive model - [RxT-Alpha-Micro-Plus](https://huggingface.co/ReactiveAI/RxT-Alpha-Micro-Plus).
-That's only a micro-scale PoC (~27M params) trained on simple synthetic datasets to demonstrate memory system. Then, we will move to bigger scales and real-world datasets in RxT-Alpha-Mini and RxT-Alpha
+> rated as ~4.36/5.0.
 
 
 ## RxNN Platform
 
 <img src="https://raw.githubusercontent.com/RxAI-dev/RxNN/refs/heads/main/assets/logo/logo_rxnn_v2.png" width="350" />
 
-We are working on complete Reactive Neural Networks development framework - [RxNN github](https://github.com/RxAI-dev/RxNN)
 
 ## Additional Research
-- **Sparse Query Attention (SQA)** - the most cost-effective GQA variant, even 2-3x faster for long sequences! Research in progress - [draft](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/sparse_query_attention.md)
+- **Sparse Query Attention (SQA)** - the most cost-effective GQA variant, even 2-3x faster for long sequences!
 - **Flex-SQA** - combination of Flex Attention and (symmetric) Sparse Query Attention, enabling 4-8x longer sliding windows
 - **Flex Memory Attention/Memory Cross-Attention** - connecting spatially sparse attention with memory layers to enable very long single interactions - smaller sliding window for input sequences attends to full memory, or the opposite
-- **Mixture-of-Experts for Grouped Attention** - MoE Router dynamically selects GQA/SQA groups, instead of static selection. Abandoned, because results were worse than for GQA/SQA  - [more](https://github.com/RxAI-dev/RxNN/blob/main/docs/research/moe_attention.md)
+- **Mixture-of-Experts for Grouped Attention** - MoE Router dynamically selects GQA/SQA groups, instead of static selection. Abandoned, because results were worse than for GQA/SQA