🎯 Reinforcement Learning from Human Feedback

🎮RL Academic

arxiv.org·

Less-relevant results

APOSM: Pairwise preference learning improves generative small-molecule design

🧠Deep Learning Academic

biorxiv.org·

Fine tuning classification in Elixir

🧠Deep Learning

elixirstatus.com·

(Mis)generalization of Helpful-Only Fine-tuning

🎮RL

lesswrong.com·

A new chapter of efficient foundation models for medical imaging

⚙️ML Systems

techcommunity.microsoft.com

Evolution of crystal field and intra-ionic interactions in ilmenite $A{\mathrm{IrO}}_{3}$ ($A=\mathrm{Mg}$, Zn, Cd) and hyperhoneycomb $β\text{−}{\mathrm{ZnIrO}...

🌐Distributed Systems

link.aps.org·

magenta/magenta-realtime: Magenta RealTime 2: An Open-Weights Live Music Model

🤖ML Code

github.com·

The Enormous Potential For Microsoft Frontier Fine Tuning

⚡Performance

joshbersin.com·

Why LLMs (still) lack taste

🎮RL

beyondtheprior.com··Hacker News

Mult-DPO: Multinomial Direct Preference Optimization for Recommender Systems

🎮RL Academic

arxiv.org·

Nvidia Nemotron 3 Ultra

🎮RL

research.nvidia.com··Hacker News

Med Tech Gurus: Why Most Radiology AI Fails

🧠Deep Learning Audio

med-tech-gurus.libsyn.com·

Vibe Diaries: Training Nanochat

🤖ML

vibediary.dev··Hacker News

Domain-Specific Small Language Models (Manning)

🤖ML

i-programmer.info·

Ideogram 4.0 launches with 2K resolution and top open-weight ranking

⚙️ML Systems

alternativeto.net·

SecLoRA: Secure Aggregation of Low-Rank Matrix Products via Functional Encryption

🌐Distributed Systems

eprint.iacr.org·

Adaptive Spo11 RNA editing gate optimizes meiosis I pace and mitotic proliferation while preserving ascospore formation

🛠️Systems Programming

science.org·

New comment by perturbation in "Ask HN: Who wants to be hired? (June 2026)"

🤖ML

drive.google.com··Hacker News

GPT-2: Too Dangerous To Release (2019)

🧠Deep Learning Blog

naokishibuya.github.io··Hacker News

Reasoning RL in 2026: GRPO, DPO, RLVR, Agentic PO & Beyond

Variational Proximal Policy Optimization

APOSM: Pairwise preference learning improves generative small-molecule design

Fine tuning classification in Elixir

(Mis)generalization of Helpful-Only Fine-tuning

A new chapter of efficient foundation models for medical imaging

Evolution of crystal field and intra-ionic interactions in ilmenite $A{\mathrm{IrO}}_{3}$ ($A=\mathrm{Mg}$, Zn, Cd) and hyperhoneycomb $β\text{−}{\mathrm{ZnIrO}...

magenta/magenta-realtime: Magenta RealTime 2: An Open-Weights Live Music Model

The Enormous Potential For Microsoft Frontier Fine Tuning

Why LLMs (still) lack taste

Mult-DPO: Multinomial Direct Preference Optimization for Recommender Systems

Nvidia Nemotron 3 Ultra

Med Tech Gurus: Why Most Radiology AI Fails

Vibe Diaries: Training Nanochat

Domain-Specific Small Language Models (Manning)

Ideogram 4.0 launches with 2K resolution and top open-weight ranking

SecLoRA: Secure Aggregation of Low-Rank Matrix Products via Functional Encryption

Adaptive Spo11 RNA editing gate optimizes meiosis I pace and mitotic proliferation while preserving ascospore formation

New comment by perturbation in "Ask HN: Who wants to be hired? (June 2026)"

GPT-2: Too Dangerous To Release (2019)