ニュース
Learning to Interpret Weight Differences in Language Models
arXiv:2510.05092v3 Announce Type: replace-cross Abstract: Finetuning (pretrained) language models is a standard approach for updating...
Learning to Align, Aligning to Learn: A Unified Approach for Self-Optimized Alignment
arXiv:2508.07750v1 Announce Type: cross Abstract: Alignment methodologies have emerged as a critical pathway for enhancing...
Learning the Wrong Lessons: Syntactic-Domain Spurious Correlations in Language Models
arXiv:2509.21155v2 Announce Type: replace Abstract: For an LLM to correctly respond to an instruction it...
Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs
arXiv:2509.22646v1 Announce Type: cross Abstract: Can humans identify AI-generated (fake) videos and provide grounded reasons?...
Learn and Unlearn: Addressing Misinformation in Multilingual LLMs
arXiv:2406.13748v3 Announce Type: replace Abstract: This paper investigates the propagation of harmful information in multilingual...
LayerNorm and RMS Norm in Transformer Models
This post is divided into five parts; they are: • Why Normalization is Needed in...
Layered Insights: Generalizable Analysis of Authorial Style by Leveraging All Transformer Layers
arXiv:2503.00958v2 Announce Type: replace Abstract: We propose a new approach for the authorship attribution task...
Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation
arXiv:2508.10404v1 Announce Type: new Abstract: With the rapid proliferation of Natural Language Processing (NLP), especially...
Layer-wise Minimal Pair Probing Reveals Contextual Grammatical-Conceptual Hierarchy in Speech Representations
arXiv:2509.15655v1 Announce Type: new Abstract: Transformer-based speech language models (SLMs) have significantly improved neural speech...
Latent Visual Reasoning
arXiv:2509.24251v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have achieved notable gains in...
Latent Fusion Jailbreak: Blending Harmful and Harmless Representations to Elicit Unsafe LLM Outputs
arXiv:2508.10029v1 Announce Type: new Abstract: Large language models (LLMs) demonstrate impressive capabilities in various language...
Large Language Models Show Signs of Alignment with Human Neurocognition During Abstract Reasoning
arXiv:2508.10057v1 Announce Type: cross Abstract: This study investigates whether large language models (LLMs) mirror human...
