We work on agentic coding, multimodal evaluation, reasoning training, low-resource NLP, and model auditing, writing each result up as a bilingual explainer for researchers and general readers alike.
arXiv · 2026
Paid search once pushed dubious hospitals in front of patients; today sellers can rewrite pages so AI favors them. SafeGEO measures the real scale of that risk — and the room for defense — across 600 recommendation cases and 22 attack variants.
AI Safety · Recommendation Agents · GEO
KDD 2026 (CCF-A) · with Xiaohongshu
50 real iOS feature tasks, 449 human-written tests, ~500K lines of production code, and a best score of 12%.
Coding Agents · Benchmark · Xiaohongshu
arXiv · 2026
Turning “remitted prior to the commencement date” into “pay before the start date” decides whether public information is readable — yet outside English it barely had an evaluation. OasisSimp has native speakers write multi-reference simplifications for 9,519 sentences in five languages, all open.
Multilingual NLP · Dataset
arXiv · 2026
Engines are like master craftsmen who cannot teach: accurate, but unable to explain. Master Distillation gets a 4B model to 48.1% puzzle accuracy, above most frontier models, with far shorter explanations.
Distillation · RLVR
arXiv · 2026
A student who never checks their paper won’t fix mistakes; neither do models trained the standard way. ThinkTwice trains solving and revision with one correctness reward, gaining 11.5 points on AIME pass@4.
RLVR · Self-refinement
From model interpretation and cross-modal evaluation to reasoning, coding agents, and AI safety, each step extends a verifiable and reproducible research program.
Readable, comparable model capability reports establish the foundation for the lab’s evaluation work.
View research9,600 evaluations quantify whether models stay consistent across text and image inputs.
View researchCoding agents, RLVR, multilingual NLP, recommendation-agent safety, and grounded reasoning.
View 2026 research