TokenSmith: Streamlining Data Editing, Search, and Inspection for Large-Scale Language Model Training and Interpretability

TL;DR

TokenSmith is an open-source tool that simplifies editing, inspecting, and analyzing large-scale training datasets for language models, enhancing workflow efficiency and accessibility.

Contribution

It introduces a modular, user-friendly library for dataset management that integrates seamlessly with existing pretraining frameworks, enabling structured data editing without changing training code.

Findings

Supports a wide range of dataset operations

Enables structured editing without code modifications

Facilitates dataset debugging and validation

Abstract

Understanding the relationship between training data and model behavior during pretraining is crucial, but existing workflows make this process cumbersome, fragmented, and often inaccessible to researchers. We present TokenSmith, an open-source library for interactive editing, inspection, and analysis of datasets used in Megatron-style pretraining frameworks such as GPT-NeoX, Megatron, and NVIDIA NeMo. TokenSmith supports a wide range of operations including searching, viewing, ingesting, exporting, inspecting, and sampling data, all accessible through a simple user interface and a modular backend. It also enables structured editing of pretraining data without requiring changes to training code, simplifying dataset debugging, validation, and experimentation. TokenSmith is designed as a plug-and-play addition to existing large language model pretraining workflows, thereby democratizing access to production-grade dataset tooling. TokenSmith is hosted on GitHub, with accompanying documentation, tutorials, and a demonstration video (available on YouTube).