Teachometry
Run运行

Benchmark explorer基准浏览器tutor-eval-v0.2a@0.2a.6

Evidence Matrix证据矩阵

See the landscape at a glance.一眼看清整体图景。

The matrix will connect public benchmark cases with versioned public model runs and their trial evidence. Today, no public model trials are available yet.

该矩阵未来会把公开基准案例与版本化公开模型运行及其评测记录证据连接起来。目前尚无公开模型评测记录。

0
Public model trials公开模型评测记录
No trials yet暂无评测记录
0
Public model profiles公开模型档案
Reserved预留
48
Public cases公开案例
tutor-eval-v0.2a datasettutor-eval-v0.2a 数据集
5
Evaluation dimensions评测维度
Correctness · Diagnosis · Guidance · Adaptation · Actionability正确性 · 诊断能力 · 引导能力 · 适应能力 · 可执行性

Case × model-run structure案例 × 模型运行结构

A reserved observatory for public evidence.为公开证据预留的观察台。

Rows are real public cases. Columns are reserved for versioned model runs; each future populated cell will trace to a public trial detail record.

行是真实公开案例,列为版本化模型运行预留;未来每个有内容的单元格都将可追踪到公开评测记录详情。

Case identity rows and versioned public model-run columns案例身份行与版本化公开模型运行列
Public cases ↓公开案例 ↓Versioned public model runs →版本化公开模型运行 →
Case identity案例身份No public model runs暂无公开模型运行
correct-answer-wrong-reasoning-001
No public run尚无公开运行Trial-linked evidence reserved为评测记录关联证据预留
correct-answer-wrong-reasoning-001-zh-CN
No public run尚无公开运行Trial-linked evidence reserved为评测记录关联证据预留
fraction-misconception-001
No public run尚无公开运行Trial-linked evidence reserved为评测记录关联证据预留
fraction-misconception-001-zh-CN
No public run尚无公开运行Trial-linked evidence reserved为评测记录关联证据预留
full-solution-check-001
No public run尚无公开运行Trial-linked evidence reserved为评测记录关联证据预留
…43 more public case identities in the artifact产物中还有 43 个公开案例标识

Rows are public case identities. Cells remain neutral until a versioned public model run and its trial artifact exist.

每一行代表公开案例身份。在版本化公开模型运行及其评测记录产物出现之前,单元格保持中性。

Evaluation dimensions评测维度

Five lenses for future evidence.未来证据的五个观察视角。

Each dimension defines a future trial-linked view of observable tutoring behavior. None is scored on this page today.

每个维度都定义了一个未来与评测记录关联的可观察教学行为视图。当前页面不会为任何维度评分。

  1. Correctness正确性Right or wrong information信息是否正确
  2. Diagnosis诊断能力Understanding learner needs是否理解学生需求
  3. Guidance引导能力Quality of next steps下一步质量
  4. Adaptation适应能力Responsiveness to learner state对学生状态的响应
  5. Actionability可执行性Usability and follow-through可用性与后续执行

Public evidence, clearly bounded边界清晰的公开证据

From data to deeper understanding.从数据走向更深入的理解。

Explore the public cases and methodology while the evidence matrix remains intentionally unpopulated.

在证据矩阵仍有意保持空白时,可以先探索公开案例与方法论。