Browsing Lost Unformed Recollections: A Benchmark for Tip-of-the-Tongue Search and Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: CH-Wang, Sky, Deshpande, Darshan, Muresan, Smaranda, Kannappan, Anand, Qian, Rebecca
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915212963086336
author CH-Wang, Sky
Deshpande, Darshan
Muresan, Smaranda
Kannappan, Anand
Qian, Rebecca
author_facet CH-Wang, Sky
Deshpande, Darshan
Muresan, Smaranda
Kannappan, Anand
Qian, Rebecca
contents We introduce Browsing Lost Unformed Recollections, a tip-of-the-tongue known-item search and reasoning benchmark for general AI assistants. BLUR introduces a set of 573 real-world validated questions that demand searching and reasoning across multi-modal and multilingual inputs, as well as proficient tool use, in order to excel on. Humans easily ace these questions (scoring on average 98%), while the best-performing system scores around 56%. To facilitate progress toward addressing this challenging and aspirational use case for general AI assistants, we release 350 questions through a public leaderboard, retain the answers to 250 of them, and have the rest as a private test set.
format Preprint
id arxiv_https___arxiv_org_abs_2503_19193
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Browsing Lost Unformed Recollections: A Benchmark for Tip-of-the-Tongue Search and Reasoning
CH-Wang, Sky
Deshpande, Darshan
Muresan, Smaranda
Kannappan, Anand
Qian, Rebecca
Artificial Intelligence
Computation and Language
Information Retrieval
Multiagent Systems
We introduce Browsing Lost Unformed Recollections, a tip-of-the-tongue known-item search and reasoning benchmark for general AI assistants. BLUR introduces a set of 573 real-world validated questions that demand searching and reasoning across multi-modal and multilingual inputs, as well as proficient tool use, in order to excel on. Humans easily ace these questions (scoring on average 98%), while the best-performing system scores around 56%. To facilitate progress toward addressing this challenging and aspirational use case for general AI assistants, we release 350 questions through a public leaderboard, retain the answers to 250 of them, and have the rest as a private test set.
title Browsing Lost Unformed Recollections: A Benchmark for Tip-of-the-Tongue Search and Reasoning
topic Artificial Intelligence
Computation and Language
Information Retrieval
Multiagent Systems
url https://arxiv.org/abs/2503.19193