Zubin's picture

Zubin

zubinb

·

AI & ML interests

None yet

Recent Activity

liked a model 3 days ago

mistralai/Mistral-Medium-3.5-128B

upvoted a collection 3 days ago

Mistral Medium 3.5

liked a model 3 days ago

dealignai/Gemma-4-31B-JANG_4M-Uncensored

View all activity

Organizations

None yet

upvoted 4 collections 3 days ago

Mistral Medium 3.5

Our first flaship models handling instruction-following, reasoning, and coding in a single set of opened-weights. • 2 items • Updated 10 days ago • 16

CRACK Abliterated Models

Uncensored JANG models with CRACK (Controlled Refusal Ablation via Calibrated Knockouts) — permanent weight-level safety removal. By dealignai. • 21 items • Updated Apr 4 • 8

JANG TurboQuantized Models

Using TurboQuant as a Quantization method • 10 items • Updated 5 days ago • 5

Custom Calibrated: Kimi - GLM - MiniMax

JANGQ Calibrated Pruned versions of Kimi k2.6 made for Macs. • 2 items • Updated 9 days ago • 1

upvoted a collection 10 days ago

DeepSeek-V4

4 items • Updated 16 days ago • 622

upvoted a collection 11 days ago

MiMo-V2.5

4 items • Updated 12 days ago • 82

upvoted a collection about 1 month ago

Gemma 4

12 items • Updated 4 days ago • 779

upvoted a collection 5 months ago

GLM-4.7

3 items • Updated Jan 19 • 65

upvoted an article 5 months ago

Article

Codex is Open Sourcing AI models

Dec 11, 2025

•

82

upvoted 5 collections 5 months ago

GPT-OSS General (4.2B to 20B)

Collection of pruned GPT-OSS models spanning 1-32 experts, maintaining general capabilities across domains while reducing computational requirements. • 29 items • Updated Aug 13, 2025 • 10

GPT-OSS Pruned Experts (4.2B-20B) [IF, Science, Math, etc.]

Complete collection of domain-specialized GPT-OSS models (1-32 experts) optimized for science, math, medicine, law, safety, and instruction following. • 8 items • Updated Aug 13, 2025 • 11

Roleplaying

Creativity at cost of context & knowledge • 5 items • Updated Dec 13, 2025 • 16

Kimi-Linear-A3B

Moonshot's experimental MoE model with Kimi Delta Attention • 3 items • Updated Jan 27 • 21

Kimi-K2

Moonshot's MoE LLMs with 1 trillion parameters, exceptional on agentic intellegence • 5 items • Updated Jan 27 • 173

upvoted a paper 5 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 59

upvoted 5 collections 5 months ago

Mistral Large 3

A state-of-the-art, open-weight, general-purpose multimodal model with a granular Mixture-of-Experts architecture. • 4 items • Updated Dec 2, 2025 • 100

💧 LFM2

LFM2 is a new generation of hybrid models, designed for on-device deployment. • 28 items • Updated Apr 8 • 154

Quantized Olmo 3

Verified models. All compatible with vLLM for very fast inference. Use the 3.1 models as they are more recent. • 23 items • Updated Dec 15, 2025 • 4

The Minitron Models and Their Teachers, Quantized

10 items • Updated Aug 26, 2024 • 1

AI PC: Text Generation

Text generation LLMs that have been validated to run on the AI PC Intel® Core™ Ultra CPU and iGPU. • 158 items • Updated 20 days ago • 15