Publications

As agents improve, users delegate increasingly complex tasks involving sequences of tool calls across multiple services. These user …

Reinforcement learning from AI feedback (RLAIF) is widely used for customizing the safety policies of large language models (LLMs) at …

Best-of-n sampling improves the accuracy of large language models (LLMs) and large reasoning models (LRMs) by generating multiple …

Many recent papers have studied the development of superforecaster-level event forecasting LLMs. While methodological problems with …

Recent reasoning models show the ability to reflect, backtrack, and self-validate their reasoning, which is crucial in spotting …

Autoregressive language models (LMs) generate one token at a time, yet human reasoning operates over higher-level abstractions - …

Despite impressive capabilities, LLMs’ successes often rely on pattern-matching behaviors, yet these are also linked to OOD …

Despite their strengths, large language models (LLMs) often fail to communicate their confidence accurately, making it difficult to …

We evaluate how well Large Language Models (LLMs) latently recall and compose facts to answer multi-hop queries like “In the year …

Despite the recent observation that large language models (LLMs) can store substantial factual knowledge, there is a limited …

Large language models (LLMs) can solve complex multi-step problems, but little is known about how these computations are implemented …

Benchmarking the performance of information retrieval (IR) is mostly conducted with a fixed set of documents (static corpora). However, …

We study whether Large Language Models (LLMs) latently perform multi-hop reasoning with complex prompts such as “The mother of …

Previous work in prompt engineering for large language models has introduced different gradient-free probability-based prompt selection …

In this paper, we present our finding that prepending a Task-Agnostic Prefix Prompt (TAPP) to the input improves the …

Pretrained Language Models (LMs) memorize a vast amount of knowledge during initial pretraining, including information that may violate …

The generative retrieval model depends solely on the information encoded in its model parameters without external memory, its …

A common practice for text retrieval is to use an encoder to map the documents and the query to a common vector space and perform a …

Language Models (LMs) become outdated as the world changes; they often fail to perform tasks requiring recent factual information which …

Large Language Models (LMs) are known to encode world knowledge in their parameters as they pretrain on a vast amount of web corpus, …

Information Extraction (IE) for semi-structured document images is often approached as a sequence tagging problem by classifying each …

In open-domain question answering (QA), retrieve-and-read mechanism has the inherent benefit of interpretability and the easiness of …

We review the EfficientQA competition from NeurIPS 2020. The competition focused on open-domain question answering (QA), where systems …

The state of the art in open-domain question answering (QA) relies on an efficient retriever that drastically reduces the search space …

Automatic speech recognition (ASR) via call is essential for various applications, including AI for contact center (AICC) services. …

Recent works in dialogue state tracking (DST) focus on an open vocabulary-based setting to resolve scalability and generalization …

Answerer in Questioner’s Mind (AQM) is an information-theoretic framework that has been recently proposed for task-oriented …