In [InstructionalFingerprint](./pubs/xu-etal-2024-fingerprinting/), we present a pilot study on LLM fingerprinting as a form of very lightweight instruction tuning. Model publisher specifies a confidential private key and implants it as an instruction backdoor that causes the LLM to generate specific text when the key is present. Results on 11 popularly-used LLMs showed that this approach is lightweight and does not affect the normal behavior of the model.
Jan 21, 2024
In [BMBI](./pubs/ma-etal-2023-mitigating-bias/), we propose to mitigate bias exhibited in QA models by observing the query instance’s influence on another instance, enabling bias mitigation with extremely low resources. With our method, bias levels in multiple bias categories can be reduced without using category-specific instance-level annotation.
Oct 1, 2023
In [STAR](./pubs/ma-etal-2023-star/), we propose to synthesize training data by structure-to-text generation using Large Language Models and we show that the generated data is even more effective than human-curated data instances to boost the low-resource event extraction performance. In the [new study](pubs/xu-etal-2023-instruction-backdoors/) about LLM's backdoor attack, we demonstrate that an attacker can inject backdoors by issuing very few malicious instructions and control model behavior through data poisoning.
May 15, 2023
We propose to [formulate biomedical relation extraction as the NLI task](pubs/xu-etal-2023-nli). We design [a new evidence retrieval technique for cross-document relation extraction](pubs/lu-etal-2023-multi).
Dec 15, 2022
In DICE, we introduce a data-efficient generative model for clinical event extraction and propose the first benchmark in the domain.
Aug 1, 2022
In Summarization as Indirect Supervision for Relation Extraction, we convert the relation extraction task into a summarization formulation.
May 1, 2022