issues Search Results · language:Dune language:Python language:JavaScript language:HTML language:Java language:JavaScript
Filter by
61.7M results
Depends on
- #85
Goal
Define auditable adjudicator decisions and correction candidates.
Acceptance criteria
- Records adjudicator, queue versions, rationale, evidence, confidence, codebook need, ...
data-schema
human-adjudication
validation
ashitaka-emishi
Depends on
- #82
- #83
Goal
Create a structured disagreement log spanning identification, semantics, domains, clusters, interpretation,
violence/obligation, agency/absence, confidence, provenance, ...
human-reliability
inter-annotator
methodology
Depends on
- #84
Goal
Prioritize disagreements for independent adjudication.
Acceptance criteria
- Includes IDs, source text, coder values, reference/model summaries, category, priority, affected ...
human-adjudication
human-reliability
pipeline
Depends on
- #81
Goal
Compare each coder and eventual adjudicated results with references without conflating this with human-human agreement.
Acceptance criteria
- Flags both-against-reference, ...
human-reliability
methodology
pipeline
Depends on
- #81
Goal
Compute identification, boundary, CMT, interpretation, agency/absence, confidence, and ambiguity agreement by case,
language, and cohort.
Acceptance criteria
- Reports positive/negative ...
human-reliability
inter-annotator
pipeline
Depends on
- #79
- #80
Goal
Normalize coder outputs while preserving every row and validation decision.
Acceptance criteria
- Completion requires two independent primary coders per reported ...
human-reliability
pipeline
validation
Depends on
- #75
Goal
Define JSON and CSV contracts for coder/cohort metadata and item-level decisions, rationales, uncertainty, and
extensible case fields.
Acceptance criteria
- Supports multiple ...
data-schema
human-reliability
validation
Depends on
- #76
- #78
Goal
Generate identification and field-agreement packets plus spreadsheet and JSON templates.
Acceptance criteria
- Preserves stable IDs and source language.
- Excludes ...
blind-coding
human-reliability
pipeline
Summary
Agent Packs is currently a useful beta surface for generating Claude-ready repo assets, but it stops too early: users
can preview/copy/download files, yet they do not get a guided path for safely ...
agent-packs
enhancement
P3
ux
madara88645
Depends on
- #75
Goal
Select defensible samples across cases, languages, phases, genres, layers, ambiguity, provenance risk, and claim
importance.
Acceptance criteria
- Includes positive cases, ...
human-reliability
inter-annotator
methodology
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.