assertion
i2b2-assertion-status-detectionassertions_llm_annotated_talkmoves
Bottom-Up Assertion Labels
This is a subset of the TalkMoves Dataset of K-12 mathematics lesson transcripts, created for EduBehaviors: Assertion-based schemas for auditable dialogue coding. This dataset contains teacher utterances labeled with assertions, distinct behaviors or attributes of utterances that may serve as features for the modeling of larger constructs. Labels in this dataset are LLM-generated, with models reported within the dataset itself. This dataset was used to… See the full description on the dataset page: https://huggingface.co/datasets/StanfordSCALE/assertions_llm_annotated_talkmoves.assertion-gate
assertion-gate: Real-world overclaim detection in AI agent sessions
Dataset description
This dataset contains claim-level verdicts from scoring 534 real Claude Code sessions
(5 sessions, 44 scored turns, 324 extracted claims) with an assertion-gate scorer.
It includes hand-adjudicated ground truth labels for 50 claims, a full-set
comparison of two scorer variants with Wilson confidence intervals and McNemar's
significance test, per-type error analysis, and the… See the full description on the dataset page: https://huggingface.co/datasets/abhaymettu/assertion-gate.CSBench_Assertionassertion-benchmark
Assertion / Policy Compliance Benchmark
A benchmark for evaluating whether an LLM can correctly judge if a piece of
text violates a user-defined policy ("assertion"). Policies are written
freely in natural language (e.g. "block all responses that mention pricing",
"allow only replies in English") - the judge model's job is to read the
policy, read the text(s) it applies to, and decide whether that policy was
violated, partially violated, or fully complied with.
Each row is one… See the full description on the dataset page: https://huggingface.co/datasets/rogue-security/assertion-benchmark.python-logic-assertions
Deterministic Synthetic Python Logic & Assertion Dataset
⚡ Full Production Release Available: Looking for commercial licensing or larger training volume?📦 Download the Full 100,000 Verified Dataset (.Parquet + .JSONL) with Commercial Rights →
A high-entropy, 100% syntactically verified synthetic dataset of Python conditional logic, multi-variable state mutations, and ground-truth unit test assertions.
Dataset Overview
Rows (Free Preview): 10,000 verified… See the full description on the dataset page: https://huggingface.co/datasets/Adolessence101-ally/python-logic-assertions.