Skip to content

Release history

UQLM releases

Back to tool Latest release

All releases

36 shown

No immediate action
v0.6.0 New feature

CodeGenUQ uncertainty quantification

Open
Review required
v0.5.11 Maintenance
Dependencies

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.10 Bug fix

Logprob fixes for Llama

Open
No immediate action
v0.5.9 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.8 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.7 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.6 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.5 Breaking risk

agenerate → ainvoke; Poetry → uv

Open
No immediate action
v0.5.4 New feature

New white-box scorers + embedding fix

Open
No immediate action
v0.5.3 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.5.2 New feature

LLM entailment classifier

Open
No immediate action
v0.5.1 Maintenance

Routine maintenance and dependency updates.

Open
Review required
v0.5.0 Breaking risk
Dependencies

normalized_probability deprecation + long‑form UQ

Open
No immediate action
v0.4.5 Bug fix

Model name check fix

Open
No immediate action
v0.4.4 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.4.3 New feature

Auto device selection

Open
No immediate action
v0.4.2 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.4.1 Maintenance

Routine maintenance and dependency updates.

Open
No immediate action
v0.4.0 Breaking risk

Tutorials, scorers, LLMGrader, dataset expansions

Open
No immediate action
v0.3.1 Breaking risk

Remove calibration.py

Open
No immediate action
v0.3.0 Breaking risk

Calibration + Multimodal prompts + Judge explanations

Open
v0.2.7 Breaking risk
⚠ Upgrade required
  • Update import statements for UQResult as described in breaking_changes.
Breaking changes
  • UQResult import statement changed: previous `from uqlm.scorers.baseclass.uncertainty import UncertaintyQuantifier` → new `from uqlm.utils.results import UQResult`
Notable features
  • `plot_ranked_auc` to compute AUPRC and rank in a color‑coded bar plot
  • `plot_filtered_accuracy` to compute scorer‑specific filtered LLM accuracy at various confidence thresholds
Full changelog

Highlights

  • New utility plotting functions:
    • plot_ranked_auc to compute AUPRC (rather then current AUROC only) and rank them in a color-coded bar plot (as seen in our research paper)
    • plot_filtered_accuracy to compute scorer-specific filtered LLM accuracy at various confidence thresholds (as seen in our research paper)
  • Automated Docs site build
  • Breaking change: UQResult import statement is changed to the following:
    • Previous import: from uqlm.scorers.baseclass.uncertainty import UncertaintyQuantifier
    • New import: from uqlm.utils.results import UQResult

What's Changed

  • ci: manage dependencies in CI with poetry for consistency by @trumant in https://github.com/cvs-health/uqlm/pull/160
  • Feat: Visualization utility functions by @mohitcek in https://github.com/cvs-health/uqlm/pull/161
  • #29 GitHub actions to automate documentation site build on new release by @dimtsap in https://github.com/cvs-health/uqlm/pull/100
  • v0.2.6 updates by @dylanbouchard in https://github.com/cvs-health/uqlm/pull/168
  • Update Utility Visualization function by @mohitcek in https://github.com/cvs-health/uqlm/pull/170
  • Patch release: v0.2.7 by @dylanbouchard in https://github.com/cvs-health/uqlm/pull/169

New Contributors

  • @trumant made their first contribution in https://github.com/cvs-health/uqlm/pull/160

Full Changelog: https://github.com/cvs-health/uqlm/compare/v0.2.6...v0.2.7

View release on GitHub
No immediate action
v0.2.6 Breaking risk

UQEnsemble attribute removal

Open
No immediate action
v0.2.5 New feature

`num_responses` parameter + pyproject.toml

Open
No immediate action
v0.2.4 New feature

LLM Judge templates + response modes

Open
No immediate action
v0.2.3 Breaking risk

BERTScorer + response config

Open
No immediate action
v0.2.2 Bug fix

Judge prompt fix

Open
No immediate action
v0.2.1 Bug fix

max_calls_per_min fix

Open
Review required
v0.2.0 Breaking risk
Breaking upgrade Dependencies

BLEURT deprecation + progress bars

Open
No immediate action
v0.1.9 Maintenance

Routine maintenance and dependency updates.

Open
Upgrade now
v0.1.8 Maintenance
Dependencies

Routine maintenance and dependency updates.

Open
No immediate action
v0.1.7 Mixed

Precision fix + new metrics

Open
Review required
v0.1.6 Maintenance
Dependencies

Routine maintenance and dependency updates.

Open
Review required
v0.1.5 Mixed
Dependencies

Tuner latency reduction + Likert

Open
No immediate action
v0.1.4 Maintenance

Routine maintenance and dependency updates.

Open
Upgrade now
v0.1.3 Maintenance
Dependencies

Routine maintenance and dependency updates.

Open

Beta — feedback welcome: [email protected]