#sampling

Nucleus Sampling: Why Your LLM Doesn’t Always Pick the “Best” Word

If you’ve read my earlier posts on temperature and top-p in this series, you already know that language models don’t just output a single…

LLM Part 8 — Token Sampling

1. Where We Left Off

বিজ্ঞানের সাজে বিজ্ঞের ফাঁকি: স্যাম্পেল সাইজের এই গোলমাল ক্ষমার অযোগ্য!

আজকাল চারদিকে একটা ভারী অদ্ভুত হিড়িক দেখা গিয়েছে। হাতে একটি অত্যাধুনিক মোবাইল ফোন, তাতে ইন্টারনেট সংযোগ এবং গুগল ফর্ম নামক একটি ডিজিটাল…

One Survey Asked Rich People Ten Times More Often Than Poor People.

Here Is How to Get the Right Answer Anyway, By Hand.

Gumbel trick teaches you to Sample From a Distribution by Adding Noise and Pointing at the Biggest…

You want your model to make a random choice — pick a word, draw a sample, roll a weighted die.

You Want to Study Rare Disasters, but They Almost Never Happen.

*Free version*

Why AI is Human? Meaning as Geometry: How Embeddings Work

Close your eyes and think of the word “dog.”

Why AI is Human? Where Randomness Actually Comes From

The Dice Inside the Machine

Understanding Statistical Distributions Through One Simple Story

Disclaimer: This poster and article were created with the assistance of AI to explore a more visual and intuitive way of learning…

[W/HOM] FearDorian X Quinn — Before You Press Play

As they outgrow past growing pains, their mental maturity proves the enduring power of their deep-rooted friendship and vibrant, playful…

The Future of Sampling Operations: Why Intelligent Routing Will Define the Next Generation of…

Many of us still evaluate operational strength based on: “How many suppliers do we have?”

AI in Sampling: The Future May Not Be Fraud Detection Alone; It May Be Intelligent Respondent…

For years, the market research industry has focused heavily on fraud prevention.

Take It With a “Sample”: Why Do Surveys Go Wrong?

When the prediction fails before the analysis even begins

Flat Earth Sequence 1.2j

by PhonicFaction (OS X 10.4.11)

PT MASUSKITA UNITED— Automatic Sampler Coal System

Automatic Sampler Coal System: Solusi Sampling Batubara Otomatis yang Akurat dan Efisien

Doggiebox 1.1.1

by Zygoat (OS X 10.4.11)

Central Limit Theorem and Normal Distribution — Case Study

— -

Probability & Sampling for Data Analysts: Understanding Uncertainty in Data

Understanding uncertainty, sampling variation, and why averages stabilize

Stop Using SMOTE Blindly: The Truth About Imbalanced Data

Why “Class Overlap” is the real enemy, why SMOTE fails in production, and how to fix it with Borderline-SMOTE and Tomek Links.

Mewakili yang Banyak dengan yang Sedikit: Populasi dan Sampel dalam Statistika

Dalam kegiatan penelitian, jumlah observasi yang dikaji dapat bervariasi: bisa berjumlah kecil, besar namun terbatas, atau bahkan tidak…

Warm Forms — Holding Hands Under the Purple Suns — Robotic Tooth, 2025

Tim Shram is quite the contributor to the Utica music scene — helpful tech at Utica DIY shows, drummer for the absolutely fantastic Trash…

LIVEN Ambient Ø Sampling Secrets: The Ultimate Guide

I created a jam session video combining SONICWARE’s ambient-focused groove box “Ambient Ø”, Bastl Instruments’ tiny modular drum synth…

GUidPlayer 1.3.3

by UNBEARABLE DEVICES (OS 9.2.2)

GABBA V1.1b

by Beatwife (OS X 10.6.8)

layers b v0.5

by ixiaudio (OS 9.2.2)

Min-p sampling for LLMs

A novel and efficient method of balanced text generation

The Importance of Sampling in Research: Validity, Bias, and Reliability

Among all methodological decisions in research, choosing the sampling method is the most crucial. It directly determines the validity…

AI vs. AI: How We Use Fake Data to Catch Real Fraud

In today’s digital world, financial fraud is growing more sophisticated, and so should our tools to prevent it. At the heart of modern…

The top_k and top_p Parameters Explained

Learn how top-k and top-p (nucleus) sampling work in text generation, with Python examples and tips for tuning these parameters.

LLM Sampling Explained: Selecting the Next Token

Learn how LLMs choose their next token using probabilities, log probabilities, greedy sampling, and random sampling — with real examples.

CV-31 (LiDAR Series 1: PointNet++)

PointNet++ applies hierarchical learning by combining: a) FPS sampling, b) Ball query grouping, c) local PointNet-style feature learning…

Geodesics, Farthest Point Sampling and Coons Patch

Dijkstra’s algorithm, sampling and basic patching

Establishing the Largest Pampers Trial Program in the Middle East: A Case Study from Egypt

Unveiling the Microbial Universe: Methods of Air Sampling for Microbiological Analysis

Quoting Olympia. The Punk Ethos at the Dawn of Modern Art

Sample Culture from Fine Art to Post Modernism

Importance Sampling Explained End-to-End

Importance sampling is a useful technique when it’s infeasible for us to sample from the real distribution p, when we want to reduce…

KLF, Sampling Pioneers, Remake a Controversial Scrapped Album and Donate It

Sampling, verbosity, and the case for (much) broader applications of distributed tracing

Originally posted as this twitter thread.

Four (+1) of the greatest songs of all time

And how these songs inspired me to create my own soundtrack

Audio sketch pad to get shit done/down

As I find myself getting out of a longer than usual musical hiatus. Thoughts of just sketching things down as they come together.