Download data/README.md from VegElpy/kitten-prompt-experiments: direct link, hf CLI and curl.
- Browser
- Download file 1.55 kB
-
https://huggingface.co/datasets/VegElpy/kitten-prompt-experiments/resolve/main/data/README.md
- Command line
-
hf download hf://datasets/VegElpy/kitten-prompt-experiments/data/README.md
-
curl -L -o README.md https://huggingface.co/datasets/VegElpy/kitten-prompt-experiments/resolve/main/data/README.md
Data tables
Each JSONL file uses one JSON object per literal-LF-delimited row. Exact decoded prompt_text, visible response_text, and provider summary text are preserved. Null means unavailable, not inferred. Per-row dates are omitted; collection occurred during August 2026.
early_prompt_variants_responses.jsonl contains 182 later first-turn variants and has no companion retrospective coding table. Source boundaries are documented in DATA_RIGHTS.md.
The five experiment response files contain 804 visible responses. The public pilot contributes 24 responses from two later adapted prompt families; its verbatim original-prompt condition is excluded. Requested, resolved, and returned model fields are distinct. suite_provider_intent records the suite design, while provider_constraint_sent: false records that the historical successful requests omitted that constraint. returned_provider records observed routing after generation.
provider_returned_reasoning_summaries.jsonl contains all 443 nonempty provider-returned summaries from the packaged pilot, replication, follow-up, and extremes rows. These are observable API-returned summaries, not raw chain of thought, complete traces, or verified causal reasoning.
Join tables with public_response_id. See the public schemas.
For ordinary-language definitions of run, sample, condition, provider, model snapshot, and reasoning summary, see the plain-language glossary.