2 5 4

Robert Dahlke PRO

rbrt

https://www.tngtech.com

AI & ML interests

None yet

Recent Activity

updated a model about 12 hours ago

tngtech/DeepSeek-R1T-Chimera

new activity 3 days ago

tngtech/DeepSeek-R1T-Chimera:Questions on how routed experts are merged

commented on their article 4 days ago

Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time

View all activity

Organizations

rbrt's activity

updated a model about 12 hours ago

tngtech/DeepSeek-R1T-Chimera

Text Generation • Updated about 12 hours ago • 2.24k • 214

New activity in tngtech/DeepSeek-R1T-Chimera 3 days ago

Questions on how routed experts are merged

#1 opened 9 days ago by

chuhac

commented on Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time 4 days ago

We published the experts that we switched off in the paper (see below). The method to switch them off works at inference time, so no need to upload new weights:

liked a model 5 days ago

deepseek-ai/DeepSeek-Prover-V2-671B

Text Generation • Updated 5 days ago • 2.03k • • 671

liked a model 8 days ago

tngtech/DeepSeek-R1T-Chimera

Text Generation • Updated about 12 hours ago • 2.24k • 214

published a model 9 days ago

tngtech/DeepSeek-R1T-Chimera

Text Generation • Updated about 12 hours ago • 2.24k • 214

updated a Space 13 days ago

README

🚀

TNG on huggingface

liked a model about 1 month ago

deepseek-ai/DeepSeek-V3-0324

Text Generation • Updated Mar 27 • 338k • • 2.85k

updated a model about 2 months ago

tngtech/OLMo-2-Instruct-Math-32B

Text Generation • Updated 19 days ago • 14 • 6

liked a model about 2 months ago

tngtech/OLMo-2-Instruct-Math-32B

Text Generation • Updated 19 days ago • 14 • 6

authored a paper 2 months ago

Mixture of Tunable Experts -- Behavior Modification of DeepSeek-R1 at Inference Time

Paper • 2502.11096 • Published Feb 16 • 2

commented 2 papers 2 months ago

Mixture of Tunable Experts -- Behavior Modification of DeepSeek-R1 at Inference Time

Paper • 2502.11096 • Published Feb 16 • 2 •

Mixture of Tunable Experts -- Behavior Modification of DeepSeek-R1 at Inference Time

Paper • 2502.11096 • Published Feb 16 • 2 •

upvoted a paper 2 months ago

Mixture of Tunable Experts -- Behavior Modification of DeepSeek-R1 at Inference Time

Paper • 2502.11096 • Published Feb 16 • 2

published a Space 2 months ago

README

🚀

TNG on huggingface

commented on Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time 2 months ago

That's a bit scary. Because listening to it, we assumed it read the paper.

So it seems there are more hallucinations than we initially thought. Funnily, quite a few of them it guessed correctly.

commented on Mixture of Tunable Experts - Behavior Modification of DeepSeek-R1 at Inference Time 2 months ago

Hey Nguyen, thank you! It's really good, actually. Congrats to your podcast generator!

Is the podcast based on the paper only? Or on blogpost + paper?

Deepseek does exaggerate a bit, though. But not too much.