Education Hub for Generative AI

Tag: tokenization

Logit Bias and Token Banning: How to Steer LLM Outputs Without Retraining 17 April 2026

Logit Bias and Token Banning: How to Steer LLM Outputs Without Retraining

Learn how to use Logit Bias and token banning to precisely steer LLM outputs, prevent unwanted words, and align brand voice without the cost of retraining.

Susannah Greenwood 7 Comments

About

AI & Machine Learning

Latest Stories

LLM Inference Observability: Tracking Token Metrics, Queues, and Tail Latency

LLM Inference Observability: Tracking Token Metrics, Queues, and Tail Latency

Categories

Featured Posts

Distributed Training at Scale: Orchestrating Thousands of GPUs for LLMs

Distributed Training at Scale: Orchestrating Thousands of GPUs for LLMs

LLM-as-a-Judge: How to Use AI Models to Evaluate Other LLMs

LLM-as-a-Judge: How to Use AI Models to Evaluate Other LLMs

Bias in Generative AI: How Data and Design Shape Outcomes

Bias in Generative AI: How Data and Design Shape Outcomes

Anonymization vs Pseudonymization in LLM Workflows: A Practical Guide

Anonymization vs Pseudonymization in LLM Workflows: A Practical Guide

Task-Specific Scorecards: Evaluating LLM Summarization, Q&A, and Extraction

Task-Specific Scorecards: Evaluating LLM Summarization, Q&A, and Extraction