Hugging Face's logo

Datasets:
RKB109
/
hybrid-semantic-search-20260728-dataset

Tasks:
Sentence Similarity
Modalities:
Text
Formats:
json
Languages:
English
Size:
< 1K
Tags:
Synthetic
semantic-search
evaluation
sentence-similarity
feature-extraction
text-ranking
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
id
stringlengths
26
26
input
stringlengths
35
72
label
stringclasses
3 values
context
stringclasses
6 values
source
stringclasses
6 values
variant
stringclasses
3 values
synthetic
bool
1 class
hybrid-semantic-search-1-1
How is the service architecture organized?
architecture
The service separates ingestion, indexing, retrieval, ranking, and API delivery.
search-01
direct
true
hybrid-semantic-search-1-2
In an operations review, How is the service architecture organized?
architecture
The service separates ingestion, indexing, retrieval, ranking, and API delivery.
search-01
operations
true
hybrid-semantic-search-1-3
For an evaluation case, How is the service architecture organized?
architecture
The service separates ingestion, indexing, retrieval, ranking, and API delivery.
search-01
evaluation
true
hybrid-semantic-search-2-2
In an operations review, What happens when the index becomes stale?
operations
Scheduled refreshes rebuild changed documents and retain the previous healthy index.
search-02
operations
true
hybrid-semantic-search-2-3
For an evaluation case, What happens when the index becomes stale?
operations
Scheduled refreshes rebuild changed documents and retain the previous healthy index.
search-02
evaluation
true
hybrid-semantic-search-3-1
How are access controls applied to results?
security
Retrieval filters documents by caller permissions before ranking.
search-03
direct
true
hybrid-semantic-search-3-3
For an evaluation case, How are access controls applied to results?
security
Retrieval filters documents by caller permissions before ranking.
search-03
evaluation
true
hybrid-semantic-search-4-1
Where is query ranking implemented?
architecture
The ranking layer combines lexical overlap, concept weights, and metadata boosts.
search-04
direct
true
hybrid-semantic-search-4-2
In an operations review, Where is query ranking implemented?
architecture
The ranking layer combines lexical overlap, concept weights, and metadata boosts.
search-04
operations
true
hybrid-semantic-search-5-1
How do operators detect poor search quality?
operations
Offline relevance sets and production click signals track quality regressions.
search-05
direct
true
hybrid-semantic-search-5-2
In an operations review, How do operators detect poor search quality?
operations
Offline relevance sets and production click signals track quality regressions.
search-05
operations
true
hybrid-semantic-search-5-3
For an evaluation case, How do operators detect poor search quality?
operations
Offline relevance sets and production click signals track quality regressions.
search-05
evaluation
true
hybrid-semantic-search-6-2
In an operations review, Can one user retrieve another team's documents?
security
Authorization filters must run before any document content is returned.
search-06
operations
true
hybrid-semantic-search-6-3
For an evaluation case, Can one user retrieve another team's documents?
security
Authorization filters must run before any document content is returned.
search-06
evaluation
true

Hybrid Semantic Search Service Synthetic Dataset

Summary

This dataset contains 14 training examples and 4 held-out examples for Enterprise search needs explainable retrieval quality when embedding APIs are unavailable, expensive, or restricted.

Every record is synthetic and includes:

  • input: query, event, or feature description
  • label: expected class, route, relation, or evidence category
  • context: synthetic supporting context
  • source: fictional source identifier
  • variant: generation pattern
  • synthetic: always true

Uses

  • Reproducible unit and integration tests
  • Baseline model training
  • Evaluation harness development
  • Schema and architecture demonstrations

Limitations

The lightweight lexical baseline is reproducible but should be replaced or compared with domain embeddings at scale.

This dataset does not represent real users, patients, customers, production traffic, or licensed media. It must not be presented as real-world evidence.

Related Model

RKB109/hybrid-semantic-search-20260728-model
Downloads last month
24
Number of rows:
18
Total file size:
11.4 kB

Models trained or fine-tuned on RKB109/hybrid-semantic-search-20260728-dataset