| 1 | paper | 11 | 12.6 | HNSW from scratch, benchmarked against FAISS: brute force still wins at 5,183 do |
| 2 | model | 10 | 11.48 | A dataset with 52 Text to image model evaluation [P] |
| 3 | models | 5 | 7.91 | A dataset with 52 Text to image model evaluation [P] |
| 4 | data | 5 | 7.75 | [R] Where does LLM document analysis/audit actually break down in production? [R |
| 5 | llm | 5 | 6.27 | [R] Where does LLM document analysis/audit actually break down in production? [R |
| 6 | etc | 5 | 5.89 | A dataset with 52 Text to image model evaluation [P] |
| 7 | open | 5 | 5.84 | Built enterprise voice intent classifier — 98.6% accuracy, 9 languages, token-fr |
| 8 | after | 3 | 5.61 | Reviewing 4 papers for AAAI 2027 and none have code, Reject? [D] |
| 9 | learning | 4 | 5.48 | Millwright — experimenting with an end-to-end machine learning framework in Rust |
| 10 | before | 5 | 5.33 | Catching bugs in scikit-learn [D] |
| 11 | benchmark | 5 | 5.07 | A dataset with 52 Text to image model evaluation [P] |
| 12 | acceptance | 3 | 4.94 | We recovered 575k crop labels from a decade of manual Photoshop work to automate |
| 13 | evaluation | 4 | 4.93 | A dataset with 52 Text to image model evaluation [P] |
| 14 | catching | 4 | 4.89 | Catching bugs in scikit-learn [D] |
| 15 | pdf | 2 | 4.89 | Continual Learning of Frontier Models for SovereignAI. Tech Report + Open Weight |
| 16 | information | 3 | 4.74 | [D] Looking for advice: Modelling a medicine-reminder agent that must decide “re |
| 17 | authors | 3 | 4.63 | Reviewing 4 papers for AAAI 2027 and none have code, Reject? [D] |
| 18 | built | 4 | 4.59 | HNSW from scratch, benchmarked against FAISS: brute force still wins at 5,183 do |
| 19 | bugs | 4 | 4.59 | Catching bugs in scikit-learn [D] |
| 20 | results | 3 | 4.55 | A dataset with 52 Text to image model evaluation [P] |