<?xml version='1.0' encoding='utf-8'?>
<rss version="2.0"><channel><title>tegridydev blog</title><link>https://tegridydev.com/blog/</link><description>Writing by tegridydev: blog</description><language>en</language><item><title>Research Papers Dataset</title><link>https://huggingface.co/blog/tegridydev/research-papers-dataset</link><description>How and why I built the research papers dataset.</description><guid isPermaLink="false">https://huggingface.co/blog/tegridydev/research-papers-dataset</guid><pubDate>Sat, 23 May 2026 00:00:00 +0000</pubDate></item><item><title>Reading LLM Model Maps: Activations and Causal Evidence</title><link>https://tegridydev.com/blog/model-interpretability/what-a-model-map-can-show/</link><description>Read a model map without confusing architecture, measured activations and causal evidence, using a local viewer and explicit capture provenance.</description><guid isPermaLink="false">urn:tegridy:blog/what-a-model-map-can-show</guid></item><item><title>Xanadu Inspired Links to Versioned Document Passages</title><link>https://tegridydev.com/blog/knowledge/xanadu-linked-documents/</link><description>Keep document versions, passage spans and link identities intact through export and import, including Unicode text and historical source references.</description><guid isPermaLink="false">urn:tegridy:blog/xanadu-linked-documents</guid></item><item><title>Organising Research Notes Without Losing History</title><link>https://tegridydev.com/blog/knowledge/sorting-notes-without-rewriting-history/</link><description>Create a useful reading copy of research notes while preserving original files, duplicate relationships, conflicting versions and attribution.</description><guid isPermaLink="false">urn:tegridy:blog/sorting-notes-without-rewriting-history</guid></item><item><title>Building Technical Knowledge Bases for Troubleshooting</title><link>https://tegridydev.com/blog/knowledge/knowledge-bases-that-help-with-real-work/</link><description>Organise a technical knowledge base around the next useful check, with source evidence, device time and receipt time kept distinct.</description><guid isPermaLink="false">urn:tegridy:blog/knowledge-bases-that-help-with-real-work</guid></item><item><title>Embedding Compatibility: Models, Vectors and Search</title><link>https://tegridydev.com/blog/knowledge/embeddings-need-a-contract/</link><description>Record model revisions, pooling and normalisation beside embeddings, then compare retrieval methods without mistaking equal dimensions for compatibility.</description><guid isPermaLink="false">urn:tegridy:blog/embeddings-need-a-contract</guid></item><item><title>A Practical Workflow for Independent AI Research</title><link>https://tegridydev.com/blog/developer-tools/research-without-losing-the-question/</link><description>Turn an AI research question into a baseline, measurable test and evidence record without letting reading lists and architecture ideas replace the question.</description><guid isPermaLink="false">urn:tegridy:blog/research-without-losing-the-question</guid></item><item><title>Synthetic Datasets with Independent Validation</title><link>https://tegridydev.com/blog/datasets/synthetic-data-generation-and-quality/</link><description>Generate synthetic examples from checkable source facts, validate answers independently and split parent records before producing related variants.</description><guid isPermaLink="false">urn:tegridy:blog/synthetic-data-generation-and-quality</guid></item><item><title>PDF to Markdown: Checking Extraction Fidelity</title><link>https://tegridydev.com/blog/datasets/pdf-extraction-and-markdown/</link><description>Check whether PDF extraction preserves page evidence, reading order and meaning, and identify when native text extraction needs an OCR alternative.</description><guid isPermaLink="false">urn:tegridy:blog/pdf-extraction-and-markdown</guid></item><item><title>Organising Research Papers, Books and Reading Notes</title><link>https://tegridydev.com/blog/datasets/paper-and-book-collection/</link><description>Keep selection, acquisition and reading separate in a research collection, with source versions and the question behind each saved paper or book.</description><guid isPermaLink="false">urn:tegridy:blog/paper-and-book-collection</guid></item><item><title>Dataset Cleaning Without Losing IDs or Meaning</title><link>https://tegridydev.com/blog/datasets/dataset-discovery-and-preparation/</link><description>Clean datasets with explicit row accounting, preserved identifiers, conflict quarantine and reproducible splits rather than silent type conversion.</description><guid isPermaLink="false">urn:tegridy:blog/dataset-discovery-and-preparation</guid></item><item><title>Extracting Contact Fields with Source Evidence</title><link>https://tegridydev.com/blog/applications/lead-extraction-and-cleaning/</link><description>Turn saved text into contact fields while keeping each value attached to its source passage and separating extraction from verification.</description><guid isPermaLink="false">urn:tegridy:blog/lead-extraction-and-cleaning</guid></item><item><title>Image Resizing: File Size, Formats and Metadata</title><link>https://tegridydev.com/blog/applications/image-conversion-resizing-and-privacy/</link><description>Choose image dimensions, fit rules and file size limits, then check the actual output and metadata instead of relying on format assumptions.</description><guid isPermaLink="false">urn:tegridy:blog/image-conversion-resizing-and-privacy</guid></item><item><title>Processing Saved Chat Logs Without Losing History</title><link>https://tegridydev.com/blog/applications/discord-log-export-and-field-extraction-toolkit/</link><description>Process saved chat JSON into traceable fields while preserving message IDs, edits, timestamps and source pages. This is an offline export workflow.</description><guid isPermaLink="false">urn:tegridy:blog/discord-log-export-and-field-extraction-toolkit</guid></item><item><title>Building a Minecraft Style Browser Game with Astra</title><link>https://tegridydev.com/blog/agents/minecraft-time-with-astra/</link><description>A personal browser game experiment with one playable original, successive screenshots and notes on the Astra and Codex revisions.</description><guid isPermaLink="false">urn:tegridy:blog/minecraft-time-with-astra</guid></item><item><title>BotSim: A Local AI Agent Community Simulator</title><link>https://tegridydev.com/blog/agents/botsim-local-community/</link><description>Explore scheduling, memory and restart behaviour in a small local agent community, with deterministic fixture checks kept separate from live model behaviour.</description><guid isPermaLink="false">urn:tegridy:blog/botsim-local-community</guid></item><item><title>Open Source AI Agents</title><link>https://huggingface.co/blog/tegridydev/open-source-ai-agents-directory</link><description>My directory of open source AI agent projects.</description><guid isPermaLink="false">https://huggingface.co/blog/tegridydev/open-source-ai-agents-directory</guid><pubDate>Fri, 21 Feb 2025 00:00:00 +0000</pubDate></item><item><title>WTF is Fine Tuning?</title><link>https://huggingface.co/blog/tegridydev/fine-tuning-dev-intro-2025</link><description>A plain English intro to fine tuning.</description><guid isPermaLink="false">https://huggingface.co/blog/tegridydev/fine-tuning-dev-intro-2025</guid><pubDate>Sun, 16 Feb 2025 00:00:00 +0000</pubDate></item><item><title>LLM Dataset Formats 101</title><link>https://huggingface.co/blog/tegridydev/llm-dataset-formats-101-hugging-face</link><description>A plain guide to common dataset formats.</description><guid isPermaLink="false">https://huggingface.co/blog/tegridydev/llm-dataset-formats-101-hugging-face</guid><pubDate>Sat, 01 Feb 2025 00:00:00 +0000</pubDate></item></channel></rss>