Presentation Rounds

Group presentations are used to develop the project in stages. Each round should turn a broad cognitive dimension into a clearer research question, evaluation plan, and final test suite.

Round 1: Cognitive Dimension and Literature

Scheduled for Week 3 and Week 4.

Focus:

Define the selected cognitive dimension.
Review classical and recent literature.
Identify representative paradigms and open questions.
Explain why the dimension matters for evaluating AI systems.

Round 2: Benchmarks, Datasets, and Replication

Scheduled for the Week 5 make-up class and Week 6.

Focus:

Identify existing SOTA benchmarks.
Identify open-source datasets.
Propose a concrete replication or evaluation workflow.
Explain what each baseline can and cannot test.

Round 3: Original Test Suite Design

Scheduled for Week 10 and Week 11.

Focus:

Propose original test cases or stimuli.
Define control conditions.
Specify target models, prompts, metrics, and expected failure modes.
Show how the new test suite goes beyond existing benchmarks.

Rubric

The presentation rubric will be posted in Handouts.