Salvato in:
Dettagli Bibliografici
Autori principali: Berdichevsky, Ruslan, Nahum-Gefen, Shai, Zaken, Elad Ben
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2510.22691
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918172355985408
author Berdichevsky, Ruslan
Nahum-Gefen, Shai
Zaken, Elad Ben
author_facet Berdichevsky, Ruslan
Nahum-Gefen, Shai
Zaken, Elad Ben
contents Despite their impressive generalization capabilities, instruction-tuned Large Language Models often underperform on text classification benchmarks. We introduce SALSA, a coherent pipeline that combines structured prompting, class-to-token mapping, and parameter-efficient fine-tuning, thereby avoiding cold-start training. Each class label is mapped to a distinct output token, and prompts are constructed to elicit a single-token response. During inference, the model's output is projected only onto the logits of the relevant class tokens, enabling efficient and accurate classification in a single forward pass. SALSA achieves state-of-the-art results across diverse benchmarks, demonstrating its robustness and scalability for LLM-based classification applications.
format Preprint
id arxiv_https___arxiv_org_abs_2510_22691
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SALSA: Single-pass Autoregressive LLM Structured Classification
Berdichevsky, Ruslan
Nahum-Gefen, Shai
Zaken, Elad Ben
Computation and Language
Machine Learning
Despite their impressive generalization capabilities, instruction-tuned Large Language Models often underperform on text classification benchmarks. We introduce SALSA, a coherent pipeline that combines structured prompting, class-to-token mapping, and parameter-efficient fine-tuning, thereby avoiding cold-start training. Each class label is mapped to a distinct output token, and prompts are constructed to elicit a single-token response. During inference, the model's output is projected only onto the logits of the relevant class tokens, enabling efficient and accurate classification in a single forward pass. SALSA achieves state-of-the-art results across diverse benchmarks, demonstrating its robustness and scalability for LLM-based classification applications.
title SALSA: Single-pass Autoregressive LLM Structured Classification
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2510.22691