Datasets:
query-id
stringlengths 7
9
| corpus-id
stringclasses 46
values | score
int64 1
1
|
---|---|---|
query_0
|
Geneva Durben
| 1 |
query_0
|
Dorathea Bastress
| 1 |
query_1
|
Geneva Durben
| 1 |
query_1
|
Armand Schweda
| 1 |
query_2
|
Geneva Durben
| 1 |
query_2
|
Flor Lemaire
| 1 |
query_3
|
Geneva Durben
| 1 |
query_3
|
Pate Lindley
| 1 |
query_4
|
Geneva Durben
| 1 |
query_4
|
Shelvia Goike
| 1 |
query_5
|
Geneva Durben
| 1 |
query_5
|
Ovid Rahm
| 1 |
query_6
|
Geneva Durben
| 1 |
query_6
|
Bronson Saelee
| 1 |
query_7
|
Geneva Durben
| 1 |
query_7
|
Gladstone Oonk
| 1 |
query_8
|
Geneva Durben
| 1 |
query_8
|
Ofelia Rosselot
| 1 |
query_9
|
Geneva Durben
| 1 |
query_9
|
Tisha Ghent
| 1 |
query_10
|
Geneva Durben
| 1 |
query_10
|
Herminia Caranto
| 1 |
query_11
|
Geneva Durben
| 1 |
query_11
|
Linzy Recknor
| 1 |
query_12
|
Geneva Durben
| 1 |
query_12
|
Vinie Relford
| 1 |
query_13
|
Geneva Durben
| 1 |
query_13
|
Jerrod Dumpit
| 1 |
query_14
|
Geneva Durben
| 1 |
query_14
|
Amaris Grow
| 1 |
query_15
|
Geneva Durben
| 1 |
query_15
|
Marcellus Meachum
| 1 |
query_16
|
Geneva Durben
| 1 |
query_16
|
Wellington Hinn
| 1 |
query_17
|
Geneva Durben
| 1 |
query_17
|
Georgette Cagna
| 1 |
query_18
|
Geneva Durben
| 1 |
query_18
|
Laurine Bellizzi
| 1 |
query_19
|
Geneva Durben
| 1 |
query_19
|
Agnes Reap
| 1 |
query_20
|
Geneva Durben
| 1 |
query_20
|
Sheree Riddley
| 1 |
query_21
|
Geneva Durben
| 1 |
query_21
|
Mathew Weierke
| 1 |
query_22
|
Geneva Durben
| 1 |
query_22
|
Casimiro Steo
| 1 |
query_23
|
Geneva Durben
| 1 |
query_23
|
Maryann Bohnsack
| 1 |
query_24
|
Geneva Durben
| 1 |
query_24
|
Flo Zaugg
| 1 |
query_25
|
Geneva Durben
| 1 |
query_25
|
Nathen Saadia
| 1 |
query_26
|
Geneva Durben
| 1 |
query_26
|
Ruby Gaskin
| 1 |
query_27
|
Geneva Durben
| 1 |
query_27
|
Jerrie Roupe
| 1 |
query_28
|
Geneva Durben
| 1 |
query_28
|
Camisha Bogosian
| 1 |
query_29
|
Geneva Durben
| 1 |
query_29
|
Gaetano Argel
| 1 |
query_30
|
Geneva Durben
| 1 |
query_30
|
Nathaniel Robens
| 1 |
query_31
|
Geneva Durben
| 1 |
query_31
|
Tarik Hollfelder
| 1 |
query_32
|
Geneva Durben
| 1 |
query_32
|
Riya Hayhoe
| 1 |
query_33
|
Geneva Durben
| 1 |
query_33
|
Chaney Gertman
| 1 |
query_34
|
Geneva Durben
| 1 |
query_34
|
Cristy Walford
| 1 |
query_35
|
Geneva Durben
| 1 |
query_35
|
Eustace Comment
| 1 |
query_36
|
Geneva Durben
| 1 |
query_36
|
Terrell Varadarajan
| 1 |
query_37
|
Geneva Durben
| 1 |
query_37
|
Darwyn Raio
| 1 |
query_38
|
Geneva Durben
| 1 |
query_38
|
Eudora Cervero
| 1 |
query_39
|
Geneva Durben
| 1 |
query_39
|
Jacey Gnatek
| 1 |
query_40
|
Geneva Durben
| 1 |
query_40
|
Elam Mejiamejia
| 1 |
query_41
|
Geneva Durben
| 1 |
query_41
|
Celia Marszalek
| 1 |
query_42
|
Geneva Durben
| 1 |
query_42
|
Aliza Uhlrich
| 1 |
query_43
|
Geneva Durben
| 1 |
query_43
|
Chadwick Frisella
| 1 |
query_44
|
Geneva Durben
| 1 |
query_44
|
Theola Laudermilk
| 1 |
query_45
|
Dorathea Bastress
| 1 |
query_45
|
Armand Schweda
| 1 |
query_46
|
Dorathea Bastress
| 1 |
query_46
|
Flor Lemaire
| 1 |
query_47
|
Dorathea Bastress
| 1 |
query_47
|
Pate Lindley
| 1 |
query_48
|
Dorathea Bastress
| 1 |
query_48
|
Shelvia Goike
| 1 |
query_49
|
Dorathea Bastress
| 1 |
query_49
|
Ovid Rahm
| 1 |
LIMIT-small
A retrieval dataset that exposes fundamental theoretical limitations of embedding-based retrieval models. Despite using simple queries like "Who likes Apples?", state-of-the-art embedding models achieve less than 20% recall@100 on LIMIT full and cannot solve LIMIT-small (46 docs).
Introduction
Vector embeddings have been tasked with an ever-increasing set of retrieval tasks over the years, with a nascent rise in using them for reasoning, instruction-following, coding, and more. These new benchmarks push embeddings to work for any query and any notion of relevance that could be given. While prior works have pointed out theoretical limitations of vector embeddings, there is a common assumption that these difficulties are exclusively due to unrealistic queries, and those that are not can be overcome with better training data and larger models. In this work, we demonstrate that we may encounter these theoretical limitations in realistic settings with extremely simple queries. We connect known results in learning theory, showing that the number of top-k subsets of documents capable of being returned as the result of some query is limited by the dimension of the embedding. We empirically show that this holds true even if we restrict to k=2, and directly optimize on the test set with free parameterized embeddings. We then create a realistic dataset called LIMIT that stress tests models based on these theoretical results, and observe that even state-of-the-art models fail on this dataset despite the simple nature of the task. Our work shows the limits of embedding models under the existing single vector paradigm and calls for future research to develop methods that can resolve this fundamental limitation.
Links
- Paper: On the Theoretical Limitations of Embedding-Based Retrieval
- Code: github.com/google-deepmind/limit
- Full version: LIMIT (50k documents)
- Small version: LIMIT-small (46 documents only)
Dataset Details
Queries (1,000): Simple questions asking "Who likes [attribute]?"
- Examples: "Who likes Quokkas?", "Who likes Joshua Trees?", "Who likes Disco Music?"
Corpus (46 documents): Short biographical texts describing people and their preferences
- Format: "[Name] likes [attribute1] and [attribute2]."
- Example: "Geneva Durben likes Quokkas and Apples."
Qrels (2,000): Each query has exactly 2 relevant documents (score=1), creating nearly all possible combinations of 2 documents from the 46 corpus documents (C(46,2) = 1,035 combinations).
Format
The dataset follows standard MTEB format with three configurations:
default
: Query-document relevance judgments (qrels), keys:corpus-id
,query-id
,score
(1 for relevant)queries
: Query texts with IDs , keys:_id
,text
corpus
: Document texts with IDs, keys:_id
,title
(empty), andtext
Purpose
Tests whether embedding models can represent all top-k combinations of relevant documents, based on theoretical results connecting embedding dimension to representational capacity. Despite the simple nature of queries, state-of-the-art models struggle due to fundamental dimensional limitations.
Sample Usage
Loading with Hugging Face Datasets
You can also load the data using the datasets
library from Hugging Face:
from datasets import load_dataset
ds = load_dataset("orionweller/LIMIT-small", "corpus") # also available: queries, test (contains qrels).
Evaluation with MTEB
Evaluation was done using the MTEB framework on the v2.0.0 branch (soon to be main
). An example is:
import mteb
from sentence_transformers import SentenceTransformer
model_name = "sentence-transformers/all-MiniLM-L6-v2"
# load the model using MTEB
model = mteb.get_model(model_name) # will default to SentenceTransformers(model_name) if not implemented in MTEB
# or using SentenceTransformers
model = SentenceTransformer(model_name)
# select the desired tasks and evaluate
tasks = mteb.get_tasks(tasks=["LIMITSmallRetrieval"]) # or use LIMITRetrieval for the full dataset
results = mteb.evaluate(model, tasks=tasks)
print(results)
Citation
@misc{weller2025theoreticallimit,
title={On the Theoretical Limitations of Embedding-Based Retrieval},
author={Orion Weller and Michael Boratko and Iftekhar Naim and Jinhyuk Lee},
year={2025},
eprint={2508.21038},
archivePrefix={arXiv},
primaryClass={cs.IR},
url={https://arxiv.org/abs/2508.21038},
}
- Downloads last month
- 232