• Updated:
  • Featured post

How LLMs get their probabilities for the next token

(este post es una explicación de la teoría. dejo aquí otro post con el detalle práctico de como usar la temperature, top_k y top_p a efectos prácticos)

Explicación más en detalle de cómo obtienen los LLMs las probabilidades para generar el siguiente token

Sampling

En cada posición el modelo tiene una bolsa con miles de tokens posibles. Sampling es el proceso de no coger siempre el token más probable, si no de orientarlo a coger respuestas con determinadas características.

Si el modelo escogiera siempre el token con más probabilidades, obtendríamos respuestas aburridas y repetitivas.

Logits

Para generar el siguiente token, una red neuronal calcula primero los vectores de logits, donde cada logit corresponde a un valor posible. El tamaño de estos vectores de logits es tan grande como el vocabulario completo del modelo.

(representación de vectores de logits)

flowchart LR
	N1["What's your favorite color?"]:::note --> Z
	Z --> A1 --> A
	Z --> B1 --> B
	Z --> C1 --> C
	Z --> D1 --> D
	
	Z["Neural network"]
    A["a"]
    A1["(-0.5)"]
    B["green"]
    B1["(0.7)"]
    C["red"]
    C1["(0.5)"]
    D["the"]
    D1["(-1.2)"]
    
    classDef note fill:none,stroke:none,color:#777;    

Los logits NO representan probabilidades ya que no suman 1 y pueden incluso ser negativos (la probabilidades no pueden). Para convertir logits a probabilidades se usa una Softmax layer

Temperature

La temperatura es una constante que se aplica a los logits antes de la transformación de la Softmax layer. Se usa para ajustar la creatividad del modelo y redistribuir la probabilidad de los valores. Una temperatura más alta hace que el modelo sea más creativo ya que aumenta las posibilidades de elegir tokens menos probables.

xychart-beta
  title "Temperatura vs Probabilidad"
  x-axis "Temperatura (T)" [0.1, 0.2, 0.5, 1, 2, 5]
  y-axis "Probabilidad" 0 --> 1
  line "P(token1)" [0.9999546, 0.9933071, 0.8807971, 0.7310586, 0.6224593, 0.5498340]
  line "P(token2)" [0.0000454, 0.0066929, 0.1192029, 0.2689414, 0.3775407, 0.4501660]

Ejemplos de temperaturas:

  • Low (0.2-0.3): El modelo es cauto y elige las palabras más probables. Output factual y predecible.
  • Medium (0.5-0.7): Un mix de confiabilidad y engagement
  • High (0.9-1.0): Toma riesgos y es impredecible

Read More

OAuth 2.0

Authentication process of verifying an identity. We confirm they’re who they say they are. (username & pwd).

Authorization process of verifying what someone is allowed to do. (Permissions and access control).

Past solutions

From worst one to best one and the problems they originate:

Credential Sharing

The worst one. An App is not able to differentiate between real user access and programmatical access.
Permissions are typically too broad. It also the ability to access more content than it should.

We could redirect the user off to the API where they could enter their credentials and get a cookie. This allows an app to access the API.

Dangerous because CSRF attacks. We’ve authorised the whole browser and not the app.

Read More

How to solve VirtualBox disk has run out of space

How to solve the problem “Low disk space on ‘Filesystem root’. The volume has only xMB disk space remaining” when you completely fill a virtual disk in VirtualBox.

(You have to delete all your snapshots first)

Open a cmd terminal and run the following command:

"c:\Program Files\Oracle\VirtualBox\VBoxManage.exe" modifymedium  
"c:\Users\mario\VirtualBox VMs\Ubuntu OTAN\Ubuntu OTAN.vdi" --resize 30000

The first path is an executable included with VirtualBox.
The second one is where your VDI actually is. --resize takes the size in MBs.

Open gpartitioner and resize it.

SCRUM PSM1 Certification - Index

psm1 badge

status: certified

This notes are my watered-down, personal version of The Scrum Guide 2020 and the following Udemy Course: “Preparation For Professional Scrum Master Level 1 (PSM1)” by Vladimir Raykov.

If you want to get ready for the certification exam, I fully recommend buying and watching his course, several times, in Udemy.

Scrum Guide 2020

Scrum Guide 2020 Notes
Scrum Glossary

“Preparation For Professional Scrum Master Level 1 (PSM1)” by Vladimir Raykov

Scrum Introduction
The Scrum Team
Scrum Events
Scrum Artifacts
Scrum Practices and Charts
A few words before the Exam
Recap of key concepts
Possible exam questions

Java Index

This are my Java-related notes. Here I have all the knowledge I refer to when I have doubts about how to use or how to implement a framework / feature I’ve already implemented once.

Version changes

Interesting changes, new functionality and APIs that come to Java with each new version. They don’t include the full changes but the ones I deemed most useful or most interesting.

From Java 8 to Java 11
Java12
Java13

Experience

Small, functional snipets on how to implement a determined feature.

Java experience sheet
How to create a database intermediate table
Java date time API
New script files in Java

Frameworks

How to use and implement determined frameworks in a Java project (using Maven).

Spring in Action (Book)
Spring Cache
Spring Beans
Thymeleaf
Spring Cors

Maven (builder)
Testing (JUnit, TestNG, Mockito)
Vert.x (microservices)
Lombok (builder)
MapStruct (mapper)

Splunk

Splunk take any type of data of millions of entries and allows you to process it into reports, dashboards and alerts.

It’s great at parsing machine data. We can train Splunk to look for certain patterns in data and label those patterns as fields.

Planning Splunk Deployments

A note on config files

Everything Splunk does is governed by configuration files. They’re stored in /etc and they’ve .conf extension.

They’re layered. You can have files with the same name in several directories. You might have a global level conf file and an app specific conf file. Splunk check which one to use based on the current app.

Read More

Oracle 1Z0-819 (Java11) Certification - Index

The new 1Z0-819 certification is the combination of the old existing ones (1Z0-815 & 1Z0-816) together.

OCP Java SE 11 Programmer I - Study guide for 1Z0-815

Welcome to Java
Java Building Blocks
Java Operators
Making Decisions
Core Java APIs
Lambdas and Functional Interfaces
Methods and Encapsulation
Class Design
Advanced Class Design
Exceptions
Java Modules

OCP Java SE 11 Programmer II - Study guide for 1Z0-816

Java Fundamentals
Java Annotations
Generics and Collections

From Java to Android with Kotlin

(Disclaimer: This are my personal notes from following Kotlin and Android courses in Udemy. This is a watered-down version from those courses. Check and buy the original courses if you want to find the full resources I used with more detail)

Android

This are my notes on the progress of things I had to learn to go from Java Developer to develop my first Android App with Android in Kotlin.

ViewBinding
DataBinding
MVVM Architecture
Live Data
ViewModel, LiveData, DataBinding
(wip: I still have to order and clean this series of posts from here on)
Recycler View
Navigation Architecture Component
Android Notifications
Coroutines
WorkManager
Android Testing

Extras:
Dagger2 Framework (dependency injection)
Hilt Framework (Dagger2 wrapper)
Room Framework (SQLite)
Android SQLite experience sheet
Android Development experience

Kotlin

This series of posts explain the main differences in language structures and usage between Kotlin and Java languages. I don’t explain the full Kotlin language, but the novelties that Kotlin implements that may be of interest to a Java developer.

From Java to Kotlin - Data Types & Casting
From Java to Kotlin - Operators & Operators Overloading
From Java to Kotlin - Nullable Types & Null Checks
From Java to Kotlin - Control Flow
From Java to Kotlin - Functions, Varargs & Default Parameters
From Java to Kotlin - Standard Library Functions
From Java to Kotlin - Lambdas
From Java to Kotlin - OOP, Companion Objects & Destructuring in Kotlin
From Java to Kotlin - Exceptions & Collections

Extras:
Kotlin cheat sheet with code examples