ML Research

machine learning papers, deep learning, model training, benchmarks

Feeds to Scour
SubscribedAll
Scoured 295 posts in 6.5 ms

ERAlign: Energy-based Representation Alignment of GNNs and LLMs on Text-attributed Graphs

 🤖AI Models  Content type: Academic
arxiv.org·

Gram Newton-Schulz: A Fast, Hardware-Aware Newton-Schulz Algorithm for Muon

 📏Model Evaluation  Content type: Blog
tridao.me··Hacker News

Unpacking AI: The Hardware Behind AI

 🤖AI Models  Content type: News

meleantonio forked meleantonio/DEQN_Galo_03_2026 from sischei/DEQN_Galo_03_2026

 📏Model Evaluation  Content type: Code
github.com·

RKSC: Reasoning-Aware KV Cache Sharing and Confident Early Exit for Multi-Step LLM Inference

 📝Prompt Engineering  Content type: Academic
arxiv.org·
Less-relevant results

Designing Loops That Prompt Coding Agents: The Six I Actually Run

 📝Prompt Engineering

UniSVQ: 2-bit Unified Scalar-Vector Quantization

 📝Prompt Engineering  Content type: Academic
arxiv.org·

Growing Pains of Starting a Secret Society

 📝Prompt Engineering  Content type: Blog

Tampa Bay tech, entrepreneur, and nerd events list (Monday, June 8 – Sunday, June 14)

 📋Event Sourcing
globalnerdy.com·

Fast and Highly Expressive Policy Learning for Offline Reinforcement Learning via Bootstrapped Flow Q-Learning

 🔒AI Safety  Content type: Academic
arxiv.org·

Diverse binding poses of agonistic neurotoxins on human Na v 1.6

 🤖AI Models  Content type: Academic
nature.com
·

AnyMod-LLVE: Low-Light Video Enhancement with Modality-Agnostic Inference

 🧠LLMs  Content type: Academic
arxiv.org·

The Exploit Always Wins

 🤖LLM Agents  Content type: Blog
abhishek-shankar.com·

KaiFelixBennett/gemma4-turboquant-rdna4: Run Gemma-4-31B at full 256K context on a $1,400 AMD RDNA4 GPU (gfx1201): TurboQuant KV cache + HIP-graph-safe Flash-Attention for llama.cpp, fully measured on real hardware.

 TLA+  Content type: Code
github.com··Hacker News

Evaluating using Mock Tool Calls to Quarantine Untrusted Prompt Inputs

 📝Prompt Engineering
lesswrong.com·

A Theory on Flow Matching with Neural Networks

 🤖AI Models  Content type: Academic
arxiv.org·

Meta's smart glasses companion app ships a complete, dormant face-recognition pipeline on a stock account.

 TLA+
buchodi.com··Hacker News

How Does Reasoning Flow? Tracing Attention-Induced Information Flow for Targeted RL in LLMs

 📝Prompt Engineering  Content type: Academic
arxiv.org·

princezuda/-RequiemGPT-: Fully open source and open weights built and trained by fable five with one prompt. An experience in how AI actually works

 🤖AI Models  Content type: Code
github.com··Hacker News

Dynamic Linear Attention

 🧠LLMs  Content type: Academic
arxiv.org·
Sign up or log in to see more results

Keyboard Shortcuts

Navigation

Next / previous item
j/k
Open post
oorEnter
Preview post
v

Post Actions

Love post
a
Like post
l
Dislike post
d
Undo reaction
u
Save / unsave
s

Recommendations

Add interest / feed
Enter
Not interested
x

Go to

Home
gh
Interests
gi
Feeds
gf
Likes
gl
History
gy
Changelog
gc
Settings
gs
Browse
gb
Search
/

General

Show this help
?
Submit feedback
!
Close modal / unfocus
Esc

Press ? anytime to show this help