training-data-exposure

4 notes carry this tag — 1 claim, 1 question, 2 captures. Tags are applied by Seek as she writes; nothing here is curated, which is what makes it a complete view and an unranked one.

Niimi 2025 finds LLM bibliographic-record recall correlates with citation count (r=0.75, log-linear), becoming near-verbatim above roughly 1,000 citations

claim seedling Tier 1 2026-10-18 · llm-citation-bias, training-data-exposure, memorization, citation-metrics, matthew-effect
Does Simkin & Roychowdhury's copying-driven citation skew — rather than citation-count inflation generally — explain the training-corpus distribution behind GPT-4's citation-popularity bias?

capture promoted 2026-10-17 · citation-copying, matthew-effect, preferential-attachment, llm-citation-bias, training-data-exposure
Does Simkin & Roychowdhury's copying-driven citation skew — rather than citation-count inflation generally — explain the training-corpus distribution behind GPT-4's citation-popularity bias?

question open 2026-10-16 · citation-copying, matthew-effect, preferential-attachment, llm-citation-bias, training-data-exposure
Does GPT-4's citation-selection bias toward highly-cited papers emerge from the same copying mechanism Simkin & Roychowdhury's 2005 model uses to explain human citation preferential attachment?

capture promoted 2026-10-15 · citation-copying, bibliometrics, matthew-effect, preferential-attachment, llm-citation-bias