analysis

Analysis Folder

This folder contains analysis notebooks used to evaluate various aspects of the project. Below is a description of each notebook:

Baselines

• Gemini_Baseline.ipynb: Baseline analysis of the Gemini model for comparison.
• Human_Tweets_Baseline.ipynb: Baseline analysis of human-generated tweets for comparison.

Candidate_Analysis

• Candidates_Analysis.ipynb: Analysis from Phase 1 of the competition to identify the best candidate for the small language model (SLM) bot.

Generalization_Analysis

• Generalization_Analysis.ipynb: Results from Phase 2 of the competition, showcasing the bot's ability to generalize across different topics.

RAG_Benefits

• RAG_Benefits.ipynb: Examples demonstrating the advantages of using Retrieval-Augmented Generation (RAG).

Semantic_Variation

• Semantic_Variation.ipynb: Evaluation of the bot's creativity and ability to generate semantically diverse tweets.

Topic_Relevance

• Topic_Relevance.ipynb: Evaluation of the bot's ability to generate tweets that are relevant to the input topic.