DKSplit on EuroHPC: Final Notes
Ten weeks have passed since we received our Playground allocation. Our GPU budget is almost spent, with fewer than 10 hours remaining. We are grateful to the EuroHPC AI Factory...
Read →EuroHPC
Model Selection Through Structured Prompting
Over the past few months, we have been refining our domain name segmentation pipeline, experimenting with a variety of models to split domain strings into meaningful words. During this process,...
Read →DKSplit on EuroHPC: Unlocking a 4B Model’s Knowledge Through Chain-of-Thought
DKSplit on EuroHPC Series #6 In our previous posts, we ran experiments across multiple architectures on EuroHPC Leonardo: BiLSTM-CRF, DeBERTa-V3, CANINE, CharBERT, ByT5-CRF, and generative LLMs. Each architecture brought incremental...
Read →DKSplit on EuroHPC: CharBERT and ByT5
In our previous update, we outlined a systematic search for a model with enough world knowledge to handle the cases that DKSplit cannot: multilingual compounds, brand portmanteaus, domain names that...
Read →From Domain Segmentation to Reading Domain Signals
Can a general-purpose open model assess brand threat without any task-specific training? We tested four model configurations on 25 public domain disputes and found that they already recognize many of the right signals. But their readings change when you change which facts are presented first.
Read →DKSplit on EuroHPC: Searching for a Teacher Model Across Architectures
Midterm report from EuroHPC Leonardo. We tested models across four architecture families. Each failed differently, and the search is narrowing.
Read →DKSplit Update: Cleaner Benchmark, First DeBERTa Run, Different Failure Modes
Continuing the work we started in our previous post. What We Did These Two Weeks Leonardo was in a planned maintenance window for most of the past two weeks. With...
Read →