Steffen Röcker's picture

Steffen Röcker PRO

sroecker

·

https://x.com/sroecker

AI & ML interests

Local models

Recent Activity

upvoted a collection 2 days ago

liked a model 4 days ago

fluxions/vui

liked a model 6 days ago

cognitivecomputations/DeepSeek-R1-0528-AWQ

View all activity

Organizations

sroecker's activity

upvoted a collection 2 days ago

MiniCPM4

MiniCPM4: Ultra-Efficient LLMs on End Devices • 17 items • Updated about 9 hours ago • 45

upvoted a collection 14 days ago

FairyR1

2 items • Updated 14 days ago • 9

upvoted a paper 14 days ago

QwenLong-L1: Towards Long-Context Large Reasoning Models with Reinforcement Learning

Paper • 2505.17667 • Published 18 days ago • 86

upvoted a paper 15 days ago

Distilling LLM Agent into Small Models with Retrieval and Code Tools

Paper • 2505.17612 • Published 18 days ago • 77

upvoted an article 25 days ago

Article

The Large Language Model Course

By

•

Jan 16

• 187

upvoted a paper 25 days ago

Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures

Paper • 2505.09343 • Published 27 days ago • 64

upvoted 3 articles about 1 month ago

Article

Good answers are not necessarily factual answers: an analysis of hallucination in leading LLMs

By

and 1 other •

May 7

• 35

Article

The 4 Things Qwen-3's Chat Template Teaches Us

By

•

Apr 30

• 51

Article

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

By

and 4 others •

Feb 18

• 33

upvoted an article about 2 months ago

Article

Tiny Agents: a MCP-powered agent in 50 lines of code

By

•

Apr 25

• 267

upvoted a collection about 2 months ago

Pleias-RAG

New generation of small reasoning models for RAG, search, and source summarization. • 4 items • Updated Apr 24 • 27

upvoted an article about 2 months ago

Article

Finetuning olmOCR to be a faithful OCR-Engine

By

and 1 other •

Apr 22

• 18

upvoted a collection about 2 months ago

Gemma 3 QAT

Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory • 15 items • Updated 11 days ago • 197

upvoted 3 collections 2 months ago

Llama 4

Meta's new Llama 4 multimodal models, Scout & Maverick. Includes Dynamic GGUFs, 16-bit & Dynamic 4-bit uploads. Run & fine-tune them with Unsloth! • 15 items • Updated 11 days ago • 46

Llama 4

Llama 4 release • 13 items • Updated Apr 29 • 526

🌙 March 2025 - Open releases from the Chinese community

32 items • Updated 25 days ago • 13

upvoted an article 3 months ago

Article

Speeding Up LLM Decoding with Advanced Universal Assisted Generation Techniques

By

and 8 others •

Mar 24

• 18

upvoted a collection 3 months ago

Gemma 3 QAT INT4 (from Flax)

These are converted from the official QAT INT4 Flax checkpoints on Kaggle. Supported formats: AutoAWQ, GGUF • 12 items • Updated Apr 6 • 5

upvoted a paper 3 months ago

SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion

Paper • 2503.11576 • Published Mar 14 • 108

upvoted a collection 3 months ago

reranking series v2

V2 crispy rerank series • 2 items • Updated Mar 13 • 23