tpi
Datasets
All datasets matching “tpi”tpi-va-corpus
TPI-VA Corpus
TPI-VA Corpus is a speech dataset for studying third-party interruption (TPI) robustness in voice assistants. A TPI setting contains a primary speaker interacting with a voice assistant and a third-party speaker who interrupts before the assistant responds. The dataset is introduced in Still Between Us? Evaluating and Improving Voice Assistant Robustness to Third-Party Interruptions.
The paper frames TPI-awareness as two linked abilities:
Discerning speaker… See the full description on the dataset page: https://huggingface.co/datasets/PleasedPenguin/tpi-va-corpus.tpips-odd-one-out
TPIPS — Odd-One-Out dataset
Image triplets with multi-factor human odd-one-out judgments. Each line of
{train,val,test}_split.jsonl is a triplet (p0, p1, p2) with a list of
factors, each carrying a soft label distribution over the three images.
Code: https://github.com/PeterWang512/TPIPS.
Images ship as images-*.tar shards (the JSONL files reference
data/odd_one_out/images/...); the download script extracts them in place.
python scripts/download.py data # downloads +… See the full description on the dataset page: https://huggingface.co/datasets/sywang/tpips-odd-one-out.tpips-2afc
TPIPS — 2AFC dataset
Two-alternative forced-choice judgments: p0 = reference, p1 = left
candidate, p2 = right candidate, with a 2-class human distribution
[left, right] per factor in annotations.jsonl.
Code: https://github.com/PeterWang512/TPIPS.
Images ship as images-*.tar shards, extracted in place by the download
script into data/2afc/ (e.g. image_editing/, nvs/, ...).
python scripts/download.py data # downloads + extracts -> data/2afc/
xP3x-tpi_Latn-Convertedclinical-tpib-pathway-stability-and-risk-guardrails-v0.1What this dataset tests
Given proposed next interventionsclassify stability in the response manifoldand add a guardrail that prevents known failure patterns.
Labels
stable_move
high_variance_move
risky_move
contraindicated_move
Typical failures
repeating tolerance loops
retrial after paradoxical worsening
allowing oscillation through exposure gaps
undertreating high-risk physiology
adding noise in flat nonresponse cases
Suggested prompt wrapper
System
You evaluate… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-tpib-pathway-stability-and-risk-guardrails-v0.1.clinical-tpib-invariant-guided-next-intervention-prediction-v0.1What this dataset tests
Given a patient’s manifold typepredict the top 3 next interventions that are most coherent.
It rewards
manifold-consistent moves
constraint-aware choices
cross-domain suggestions when warranted
It penalizes
repeating tolerance loops
repeating paradoxical worseners
ignoring contraindications
choosing common care without manifold fit
Labels
coherent_top3
partially_coherent_top3
incoherent_top3
Suggested prompt wrapper
System
You propose the… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/clinical-tpib-invariant-guided-next-intervention-prediction-v0.1.