How RAG Changes
Local LLM Output Quality
A Comparative Benchmark

Episodes

1. 01

   RAG Impact on Local LLM Output Quality - Episode 1

   Built a Gemma3-based local pipeline and compared output quality between RAG-enabled and non-RAG generations.

2. 02

   RAG Impact on Local LLM Output Quality - Episode 2 (Tokyo Game Show Prompt)

   Automated comparison report of Gemma3 outputs with and without RAG for the same topic prompt.

3. 03

   RAG Impact on Local LLM Output Quality - Episode 3 (Gemma3 Query Rewrite)

   Comparison focused on switching the query-rewrite phase to Gemma3 in the benchmark pipeline.

4. 04

   RAG Impact on Local LLM Output Quality - Episode 4 (Vectorization from Wikipedia Dumps)

   Benchmark focused on vectorization flow derived from Wikipedia dump data.

5. 05

   RAG Impact on Local LLM Output Quality - Episode 5 (MCP Server Build Guide)

   Benchmark article with a practical MCP server implementation scenario.

6. 06

   RAG Impact on Local LLM Output Quality - Episode 6 (Model Selection Strategy)

   Comparison centered on implementation patterns for dynamic LLM model selection.

7. 07

   RAG Impact on Local LLM Output Quality - Episode 7 (Bun and Biome)

   Benchmark article using Bun and Biome as the explanatory target domain.

8. 08

   RAG Impact on Local LLM Output Quality - Episode 8 (Wikipedia API Usage)

   Comparison focused on outputs for prompts about Wikipedia API usage patterns.