@@ -7935,18 +7935,62 @@ <h4 id="Assumptions-about-equipment-and-dependencies">Assumptions about equipmen
79357935</ tr >
79367936< tr >
79377937< td > < code > sentence_transformers</ code > </ td >
7938- < td > Word embeddings for novelty / surprise</ td >
7938+ < td > Word embeddings for novelty / surprise (downloads < code > thenlper/gte-large </ code > on first use) </ td >
79397939</ tr >
79407940< tr >
79417941< td > < code > benepar</ code > </ td >
7942- < td > Constituency parsing (optional pipeline extension)</ td >
7942+ < td > Constituency parsing (spaCy pipeline extension used by creativity scoring)</ td >
7943+ </ tr >
7944+ < tr >
7945+ < td > < code > numpy</ code > </ td >
7946+ < td > Numerical helpers in scoring</ td >
7947+ </ tr >
7948+ < tr >
7949+ < td > < code > streamlit</ code > </ td >
7950+ < td > Web UI (< code > streamlit_app.py</ code > )</ td >
7951+ </ tr >
7952+ < tr >
7953+ < td > < code > altair</ code > </ td >
7954+ < td > Charts in the Streamlit app</ td >
7955+ </ tr >
7956+ </ tbody >
7957+ </ table >
7958+ < p > < strong > Optional</ strong > (< code > [project.optional-dependencies]</ code > ):</ p >
7959+ < table >
7960+ < thead >
7961+ < tr >
7962+ < th > Package</ th >
7963+ < th > Role</ th >
7964+ </ tr >
7965+ </ thead >
7966+ < tbody >
7967+ < tr >
7968+ < td > < code > pytest</ code > </ td >
7969+ < td > Unit tests (< code > pip install -e ".[dev]"</ code > )</ td >
7970+ </ tr >
7971+ </ tbody >
7972+ </ table >
7973+ < p > < strong > Vignette / notebook only</ strong > (not in < code > pyproject.toml</ code > ; install if you run §3 plotting cells):</ p >
7974+ < table >
7975+ < thead >
7976+ < tr >
7977+ < th > Package</ th >
7978+ < th > Role</ th >
7979+ </ tr >
7980+ </ thead >
7981+ < tbody >
7982+ < tr >
7983+ < td > < code > matplotlib</ code > </ td >
7984+ < td > Example figures in this report</ td >
79437985</ tr >
79447986</ tbody >
79457987</ table >
79467988< p > Install from the repository root:</ p >
79477989< div class ="highlight "> < pre > < span > </ span > python< span class ="w "> </ span > -m< span class ="w "> </ span > venv< span class ="w "> </ span > .venv
79487990< span class ="nb "> source</ span > < span class ="w "> </ span > .venv/bin/activate< span class ="w "> </ span > < span class ="c1 "> # Windows: .venv\Scripts\activate</ span >
7949- pip< span class ="w "> </ span > install< span class ="w "> </ span > -e< span class ="w "> </ span > < span class ="s2 "> ".[dev]"</ span >
7991+ pip< span class ="w "> </ span > install< span class ="w "> </ span > -e< span class ="w "> </ span > < span class ="s2 "> ".[dev]"</ span > < span class ="w "> </ span > < span class ="c1 "> # package + pytest</ span >
7992+ < span class ="c1 "> # Optional, for §3 matplotlib examples:</ span >
7993+ pip< span class ="w "> </ span > install< span class ="w "> </ span > matplotlib
79507994</ pre > </ div >
79517995< h4 id ="Scenario "> Scenario< a class ="anchor-link " href ="#Scenario "> ¶</ a > </ h4 > < blockquote >
79527996< p > < em > A researcher is comparing two OpenRouter models on cognitive flexibility and creativity. They clone the miniBen repo, add their API key to < code > .env</ code > , list available benchmarks from the terminal, then run the creativity benchmark for each model. miniBen makes five API calls per model, parses each story, computes per-call surprise and aggregate < code > avg_novelty</ code > , and prints a summary they can paste into their lab report.</ em > </ p >
0 commit comments