Metraipolis › archive › dc_387bb2d3f36e
Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification — arXiv:2608.13866
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13866
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T02:28:00.497Z
- SHA-256 of the archived text
c1ec279e9c3e04a5848d2086b3e87cde042d8e9c4a27c33024208343f66fd33d- Stable URL of this page
https://metraipolis.com/documento/dc_387bb2d3f36e
How to cite
Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification — arXiv:2608.13866 — arXiv (en), https://arxiv.org/abs/2608.13866 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_387bb2d3f36e (Metraipolis archive dc_387bb2d3f36e, sha256 c1ec279e9c3e04a5848d2086b3e87cde042d8e9c4a27c33024208343f66fd33d)
BibTeX
@misc{dc387bb2d3f36e,
title = {Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification — arXiv:2608.13866},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13866},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_387bb2d3f36e; sha256 c1ec279e9c3e04a5848d2086b3e87cde042d8e9c4a27c33024208343f66fd33d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13866. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) can generate synthetic training data for text classification, but the quality of generated samples is heterogeneous: some fall in correct class regions of the embedding space while others land in peripheral or cross-class zones. We propose a geometric filtering framework that evaluates each LLM-generated sample by its Euclidean distance to real class examples in a sentence embedding space, selecting only geometrically consistent candidates. A soft weighting mechanism transforms filter scores into sample weights for classifier training. Evaluated across 13 datasets, 5 classifiers, 10 augmentation methods, and over 6,700 configurations, our method achieves +2.61 percentage points (pp) over SMOTE ($p<0.0001$, Cohen's $d=0.95$, 88.9% win rate). The approach generalizes to named entity recognition (+9.26pp, 100% win rate) without filter modification, and is robust across 5 LLMs from 4 providers. A key finding is that the simplest distance-based filter consistently outperforms complex multi-criteria alternatives. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-17 in cs.CL). Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.13866 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-19T02:20:31.917Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).