Posts by Collection

projects

Archively

Dates: – Present | Location: CITP @ Princeton University

Collaborators: Marianne Aubin Le Quere, Manoel Horta Ribeiro, Andy Guess, Tiffany Liu

Lead developer on Archively, an iOS app designed to support a study on digital well being by directing users to EMAs after completing Tiktok sessions and allowing users to upload their de-identified exported Tiktok data at the start and end of the study. Archively is built with a Flutter frontend and a FastAPI backend that utilizes Google Cloud services, such as Firebase Cloud Messaging, Data Store, and Cloud Run.

Unscrambled: An EMA Platform for Studying Time Perception in Short-Form Video

Dates: – Present | Location: Princeton University

Collaborators: Manoel Horta Ribeiro

Designed a single-mechanic ecological momentary assessment platform collapsing measurement and intervention into a guess-then-feedback task on cumulative session time, delivered via iOS Live Activities.

Counterfactual Comments: A Cross-Platform Causal Design for Comment Toxicity

Dates: – Present | Location: Princeton University

Collaborators: Mehedi Zaman, Francesco Salvi, Jason Greenfield, Manoel Horta Ribeiro

This project investigates the causal effect of early comment-thread toxicity on subsequent toxicity, using a novel identification strategy to isolate toxicity spillover from content-level confounding.

Pastcasting: Forecasting Historical Events with a Vintage LLM

Dates: – Present | Location: Princeton University

Collaborators: Liv d'Aliberti, Manoel Horta Ribeiro

Fine-tuned a 13B model with a 1931 training cutoff to test whether LLM forecasting accuracy degrades with temporal distance from training data.

publications

CrossWeave: Bridging Perspectives Across Online Communities with a Dual-Pane Design

Published in CSCW Companion 2026, 2026

CrossWeave is an AI-augmented social media interface that surfaces relevant posts from other conversation threads in a side pane, using LLM-identified highlights to help users discover cross-community perspectives and draft more constructive replies.

Fang, Fei, Hirave, Reva, Jurayj, William, Li, Yuqi, Lu, Brian, Metin, Tarik, Miyake, Tsugunobu, Morhun, Kateryna, Permalla, Yash, Rustamov, Kenan, Shen, Allen, Shi, Haojun, Singh, Piccardi, Tiziano, Prabhav, Wang, Xiheng Tom, Xu, Kevin, Zeng, Qingcheng, Khashabi, Daniel, Perrin, Andrew J., Xiao, Ziang, Zhang, Jiayi, and Eisner, Jason M. (2026). "CrossWeave: Bridging Perspectives Across Online Communities with a Dual-Pane Design." CSCW Companion 2026. (Authors listed alphabetically.)
Download Paper |

Who Let the Dogs Out? Auditing Dogwhistle Coverage in Hate Speech Benchmarks Across Target Groups

Published in WOAH @ EMNLP '26, 2026

Audited dogwhistle coverage across two widely-used hate speech benchmarks (HateXplain, Measuring Hate Speech), mapping an existing 340-entry dogwhistle glossary onto a four-level coding sophistication taxonomy across 59.6k posts.

Hirave, Reva and Oet, Ryan (2026). "Who Let the Dogs Out? Auditing Dogwhistle Coverage in Hate Speech Benchmarks Across Target Groups." WOAH @ EMNLP '26.
Download Paper |

talks

Ongoing Work

Published:

Presented ongoing projects: Archively, Unscrambled, Pastcasting, and Benchmarking Dogwhistles

teaching