Hugging Face
Models
Datasets
Spaces
Community
Docs
Enterprise
Pricing
Log In
Sign Up
In a Training Loop 🔄
126.8
TFLOPS
4
47
78
Karsten Kuhnke
PRO
mindchain
Follow
simeks18's profile picture
webxos's profile picture
Goenzilla's profile picture
10 followers
·
75 following
https://www.linkedin.com/in/jankarstenkuhnke/
KarstenKuh16443
haddock-development
karsten-kuhnke-0b23023a3
AI & ML interests
Mechanistic Interpretability, Sparse Autoencoders, JumpReLU, Reward Modeling, RLHF, AI Alignment, Function Calling, Gemma, Nemotron
Recent Activity
reacted
to
their
post
with 🧠
about 2 hours ago
The Architecture of 2026: Beyond the Token Trap 🚀 We are witnessing a tectonic shift in Transformer architecture. It’s no longer just about "predicting the next token"—it’s about executing latent plans on a high-speed data highway. What happens when we combine DeepSeek’s stability with Google’s strategic intelligence? 1️⃣ The Infrastructure: DeepSeek’s mHC Moving from a single-lane residual stream to a multi-lane highway. Using the Birkhoff Polytope, mHC ensures mathematical stability (Identity Mapping) while routing specialized data through dedicated lanes. 2️⃣ The Intelligence: Google’s Meta-Controller An internal AI unit that lives inside the Transformer. It escapes the "Token Trap" by extracting data to create a latent plan, steering the model via Temporal Abstraction. The Synergy: In a Topological Transformer, the Meta-Controller finally has the "dedicated lanes" it needs to steer complex reasoning without causing gradient explosions. We aren't just making models bigger; we are making them architecturally smarter. 🧠 #MachineLearning #DeepSeek #GoogleAI #Transformer #AIArchitecture
upvoted
a
paper
about 4 hours ago
DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Models
updated
a collection
about 11 hours ago
RLM - Neuro-Symbolic Architecture - Reasonig Traces
View all activity
Organizations
mindchain
's models
45
Sort: Recently updated
mindchain/llama2-adapter_01
Updated
Sep 11, 2023
•
1
mindchain/llama2-13
Updated
Sep 11, 2023
mindchain/llama-2-7b-custom1
Updated
Sep 11, 2023
•
1
mindchain/llama2-7-segeln_003_adapter
Updated
Sep 11, 2023
mindchain/llama2-7-segeln_001
Updated
Sep 11, 2023
•
1
mindchain/llama2-ooiiioo
Updated
Sep 10, 2023
mindchain/llama-2-7b-custom004444444
Text Generation
•
Updated
Sep 10, 2023
•
10
mindchain/llama2-qlora-finetunined-french_aktuell
Updated
Sep 10, 2023
mindchain/llama-2-7b-custom
Updated
Sep 10, 2023
mindchain/llama2-qlora-finetunined_01
Updated
Sep 9, 2023
mindchain/llama2-qlora-finetunined-french
Updated
Sep 9, 2023
mindchain/ybelkada-falcon-7b-sharded-bf16-yizhongw_self_instruct
Updated
Jul 13, 2023
mindchain/ybelkada-falcon-7b-sharded-bf16-qlora-openassistant-guanaco
Updated
Jul 6, 2023
•
3
mindchain/t5-small-finetuned-xsum-finetuned-xsum2
Updated
May 25, 2023
•
5
mindchain/t5-small-finetuned-xsum
Updated
May 25, 2023
•
10
Previous
1
2
Next