Skip to content

Meta-Cognitive Analysis: Evaluating Declarative and ProceduralKnowledge in Datasets and Large Language Models

Notifications You must be signed in to change notification settings

Li-Z-Q/meta-cognitive-analysis

Repository files navigation

meta-cognitive-analysis

Meta-Cognitive Analysis: Evaluating Declarative and Procedural Knowledge in Datasets and Large Language Models

Prepare Experimental Datasets

python generate_raw_datas.py

Generate Hint via GPT4 and GPT3.5

# Generate Hint by In-Context Learning
python generate_hint.py
# Generate Some Noisy Factual Hint for MATH and GSM8k
python generate_noise.py

Evaluate by Hint Injection

# Include No Hint, Factual Hint, Procedural Hint, Both Factual and Procedural Hint
bash run.sh

Analyze Results

# Annotate via Some Rules
python annotate.py
# Show Improvement Scores from Hint Injection
python analysis.py

About

Meta-Cognitive Analysis: Evaluating Declarative and ProceduralKnowledge in Datasets and Large Language Models

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published