docs: add LLM fixtures for demo dataset (fine-tuning dataset), fix demo notebook

[cjunkin]

resolves #4285 and #4247

Commit "demo llm fixtures, fix notebook" is the most recent significant commit. Rebase added previous commits that were in this PR: #4202

Fixtures found here (described in detail below): https://console.cloud.google.com/storage/browser/arize-phoenix-assets/traces

Demo fixtures

• Demo RAG traces and evals (demo_llama_index_rag_traces.parquet, demo_llama_index_rag_<name for eval>.eval.parquet) for demo fixtures: https://storage.googleapis.com/arize-phoenix-assets/traces/demo_llama_index_rag_traces.parquet

Demo LLM fixtures

• Demo LLM traces: demo_llama_index_llm_.<llm type or all llm traces>.parquet
  • Two types of LLM spans exist so I uploaded fixtures that have both of them combined and separated just in case:
    LLM.predict (parent) -> OpenAI.chat (child)

Notebook-related fixtures

• Demo RAG questions fixture to run the notebook (demo_llama_index_rag_questions.parquet): https://storage.googleapis.com/arize-phoenix-assets/demos/demo_llama_index_rag_questions.parquet

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB