Evaluating and Inducing Personality in Pre-trained Language Models (2022 · 168 citations)
This NeurIPS 2023 paper asks whether personality psychology can provide a principled, quantitative way to describe how large language models behave, and whether a chosen personality can be deliberately instilled in them.
GRICE: A Grammar-based Dataset for Recovering Implicature and Conversational rEasoning (2021 · 36 citations)
GRICE is a synthetic, bAbI-style dialogue dataset. In a small simulated world of agents, objects and rooms, Alice asks questions, and Bob answers with replies that carry one of five implicature types. Relevance: «Did you see the apples?» — «There is a basket in the dining room». Strengthening: «All of them are…
Xiezhi: An Ever-Updating Benchmark for Holistic Domain Knowledge Evaluation (2023 · 34 citations)
Benchmarks such as MMLU, BIG-bench and HELM quickly become too easy for modern LLMs and may leak into their training data.
Code over Words: Overcoming Semantic Inertia via Code-Grounded Reasoning (2026)
Large language models often cannot set aside what they learned in pre-training when the context explicitly redefines how the world works.
Rational Communication Shapes Morphological Composition (2026)
The paper asks why English settled on one particular morpheme combination for a new word when other combinations, built from morphemes that existed at the time, could have named the same concept.

Co-authors here

11 co-authors with a page, the most papers together first.

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

Papers of our library cite their works 152 times (of the 149,439 papers whose bibliographies we read); 46 citing papers with a page here →.

Updated 2 November 2026