paper-with-me

홈 › Papers

Beyond Perplexity: A Lightweight Benchmark for Knowledge Retention in Supervised Fine-Tuning

2026-01-07 · Soheil Zibakhsh Shabgahi, Pedram Aghazadeh, Farinaz Koushanfar arxiv

Supervised Fine-Tuning (SFT) is a standard approach for injecting domain knowledge into Large Language Models (LLMs). However, relying on validation perplexity to monitor training is often insufficient, as it confounds stylistic mimicry with genuine factual internalization. To address this, we introduce the Knowledge Retention (KR) Test , a lightweight, corpus-grounded evaluation framework designed to distinguish factual learning from linguistics. KR-Test utilizes automatically generated contrastive examples to measure likelihood preferences for correct versus incorrect continuations, requiring no instruction tuning or generative decoding. We validate the framework's integrity through a "blind vs. oracle" baseline analysis. Furthermore, we demonstrate the diagnostic capabilities of KR-Test by analyzing the training dynamics of Low-Rank Adaptation (LoRA). By exposing the fine-grained dissociation between linguistic convergence and knowledge retention, KR-Test enhances the interpretability of fine-tuning dynamics.

📄 PDF Abstract BibTeX arXiv:2601.03505

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Beyond Perplexity: Let the Reader Select Retrieval Summaries via Spectrum Projection Score

2025-08-08 · Zhanghao Hu, Qinglin Zhu, Siya Qi, Yulan He 외 arxiv

Large Language Models (LLMs) have shown improved generation performance through retrieval-augmented generation (RAG) following the retriever-reader paradigm, which supplements model inputs with externally retrieved knowl…

The Asymmetric Harms of LLM Compression

2026-08-20 · Yuan Wu, Mairui Li, Lesia Semenova, Chudi Zhong arxiv

Large language models (LLMs) compression reduces deployment costs, but standard aggregate metrics like perplexity and accuracy often mask underlying behavioral shifts. In this work, we systematically evaluate 3 LLMs acro…

When to Review: Spaced Repetition for Continual Pre-Training of Language Models

2026-08-18 · Alankar Atreya, Devesh Batra, Yoages Kumar Mantri, Geremy Bantug 외 arxiv

Continual pre-training of large language models must acquire new information without erasing old knowledge. Existing replay methods often choose a global old/new mixture and sample uniformly, ignoring that examples diffe…

Continual Learning

Detecting Fluent Optimization-Based Adversarial Prompts via Sequential Entropy Changes

2026-05-19 · Mohammed Alshaalan, Miguel R. D. Rodrigues arxiv

Optimization-based adversarial suffixes can jailbreak aligned large language models (LLMs) while remaining fluent, weakening static and windowed perplexity-based detectors. We cast adversarial suffix detection as an onli…

Null Space Constrained Contrastive Visual Forgetting for MLLM Unlearning

2026-05-07 · Yuhang Wang, Zhenxing Niu, Haoxuan Ji, Guangyu He 외 arxiv

The core challenge of machine unlearning is to strike a balance between target knowledge removal and non-target knowledge retention. In the context of Multimodal Large Language Models (MLLMs), this challenge becomes even…