Work
PROJECTS / WRITINGS
- 2026
Small-Corpus-Guided Steering of Base Model Generation
Using RAG and activation steering to measurably shift a base LLM's generative register toward a 58-text corpus of Taiwanese queer literature, with no fine-tuning required.
→ - 2026
General Map of the Literary Encounter
A conceptual map charting the full architecture of what constitutes a literary encounter, medium, text, intertextuality, reader, the encounter, and emergence, as a scaffold for a federated computational approach to literature.
→ - 2026
Reading Stance Parameterized: LLM Agents as Proxies for the Literary Encounter
Parameterizing LLM agents with contrastive reading stances to test whether differently oriented readers produce measurably different readings of the same text.
→ - 2025
Computational Analysis on Machine Consciousness in Science Fiction
Building and examining a 115-text full-text science fiction corpus with Topic Modeling, SBERT Embeddings + UMAP Clustering, Zero-Shot Classification, and Emotion Detection.
→ - 2025
LLM Style Observatory
An interactive dashboard comparing stylometric signatures across GPT-4o, Claude, Gemini, and local models, asking whether LLMs carry stable, architecturally determined writing styles.
→ - 2025
Open Research Assistant
A multi-source research tool that aggregates open archives and academic APIs to find where a text's full version actually lives, with an optional LLM-generated reading list mode.
→ - 2019–
Textures
A personal black-and-white archive of surfaces gathered across geographies and time.
→ - 2025
Coffee Archive
A filterable archive of single-origin tasting notes drawn across roaster collections, letting you explore tendencies by country, process, varietal, and altitude.
→ - 2024
Material, ĠConfig, urations, Ġof, ĠAI, -, Gener, ated, ĠText, s
A new materialist account of how AI-generated text emerges — examining GPT-2's tokenization, embeddings, and attention mechanisms to argue that meaning in AI text is not absent but differently configured.
→ - 2024
An AI Image of the Imagined Image of AI
What does it mean that generative AI consistently imagines itself as a feminized humanoid robot? A critical examination of visual repetition, historical imaginary, and political foreclosure in AI-generated images.
→ - 2024
mOther as the Other: Language, Alienation, and the Impossibility of the Mother Tongue
A theoretical memoir examining the mother tongue through Lacanian psychoanalysis, tracing the linguistic histories of Canetti, Lahiri, and my own experience as a Taiwanese speaker.
→