datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
finephrase-minipilotfinephrase-gov-reranker-dataset-v2-200
Finephrase Gov Reranker Dataset v2 (200 rows)
Synthetic reranking dataset built from HuggingFaceFW/finephrase FAQ-style questions filtered to Irish public-service domains:
gov.ie
citizensinformation.ie
hse.ie
oireachtas.ie
revenue.ie
Each grouped row contains a query plus one relevant page snippet and a set of hard negatives from the same filtered corpus.
Included Files
train.jsonl, valid.jsonl, test.jsonl
grouped reranker evaluation/training format… See the full description on the dataset page: https://huggingface.co/datasets/temsa/finephrase-gov-reranker-dataset-v2-200.finephrase-itafinephrase-gov-reranker-hard-k10-holdout-a03
Finephrase Gov Reranker Hard K10 Holdout A03
Derived hard in-domain reranker dataset built from HuggingFaceFW/finephrase and filtered to Irish public-service domains.
This split was created as a fresh unseen overfit gate for the K=10 reranker cycle, not as the main training mix.
Intended use
fresh hard holdout for reranker overfit checks
secondary in-domain evaluation for small CPU-first rerankers
not the only release gate; use it together with bilingual… See the full description on the dataset page: https://huggingface.co/datasets/temsa/finephrase-gov-reranker-hard-k10-holdout-a03.finephrase-gov-reranker-hard-k10-v1
FinePhrase Gov Reranker Hard K10 v1
Hard reranker dataset derived from HuggingFaceFW/finephrase for Irish-government-like search.
Each row is a grouped reranker example:
one query
one relevant passage from a public-sector page
nine hard negatives from the same filtered gov-like domain pool
The candidate set is fixed at 10, matching the current serving target of reranking only the top 10 candidates after first-stage retrieval / RRF.
Source filtering
Rows are… See the full description on the dataset page: https://huggingface.co/datasets/temsa/finephrase-gov-reranker-hard-k10-v1.finephrase_testfinephrase-sample-1m