Benchmark for Assessing Olfactory Perception of Large Language Models

Develop an evaluation suite for odor reasoning tasks to test LLM sensory grounding. This allows for benchmarking models on non-textual human sensory experience.

Suggested repo: olfact-eval

"Can your LLM actually smell? A benchmark for olfactory reasoning."

Estimated effort: 20h