Image de couverture de livre

Disentangling Geometry, Performance, and Training in Language Models

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Kulkarni, Atharva, Springer, Jacob Mitchell, Subramonian, Arjun, Swayamdipta, Swabha
Format:	Preprint
Publié:	2026
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Annotating FrameNet via Structure-Conditioned Language Generation
par: Cui, Xinyue, et autres
Publié: (2024)

How Reliable is Language Model Micro-Benchmarking?
par: Yauney, Gregory, et autres
Publié: (2025)

Compare without Despair: Reliable Preference Evaluation with Generation Separability
par: Ghosh, Sayan, et autres
Publié: (2024)

Improving Language Model Personas via Rationalization with Psychological Scaffolds
par: Joshi, Brihi, et autres
Publié: (2025)

Understanding Dataset Difficulty with $\mathcal{V}$-Usable Information
par: Ethayarajh, Kawin, et autres
Publié: (2021)

Better Language Model Inversion by Compactly Representing Next-Token Distributions
par: Nazir, Murtaza, et autres
Publié: (2025)

Stop! In the Name of Flaws: Disentangling Personal Names and Sociodemographic Attributes in NLP
par: Gautam, Vagrant, et autres
Publié: (2024)

Crowd-Calibrator: Can Annotator Disagreement Inform Calibration in Subjective Tasks?
par: Khurana, Urja, et autres
Publié: (2024)

Evaluation Under Imperfect Benchmarks and Ratings: A Case Study in Text Simplification
par: Liu, Joseph, et autres
Publié: (2025)

Robust Data Watermarking in Language Models by Injecting Fictitious Knowledge
par: Cui, Xinyue, et autres
Publié: (2025)

Believing without Seeing: Quality Scores for Contextualizing Vision-Language Model Explanations
par: He, Keyu, et autres
Publié: (2025)

Teaching Models to Understand (but not Generate) High-risk Data
par: Wang, Ryan, et autres
Publié: (2025)

Logits of API-Protected LLMs Leak Proprietary Information
par: Finlayson, Matthew, et autres
Publié: (2024)

Evaluating Evaluation Metrics -- The Mirage of Hallucination Detection
par: Kulkarni, Atharva, et autres
Publié: (2025)

BenchBrowser: Retrieving Evidence for Evaluating Benchmark Validity
par: Diddee, Harshita, et autres
Publié: (2026)

Sample, Align, Synthesize: Graph-Based Response Synthesis with ConGrs
par: Ghosh, Sayan, et autres
Publié: (2025)

ELI-Why: Evaluating the Pedagogical Utility of Language Model Explanations
par: Joshi, Brihi, et autres
Publié: (2025)

NeuroComparatives: Neuro-Symbolic Distillation of Comparative Knowledge
par: Howard, Phillip, et autres
Publié: (2023)

Understanding Catastrophic Forgetting in Language Models via Implicit Inference
par: Kotha, Suhas, et autres
Publié: (2023)

Understanding "Democratization" in NLP and ML Research
par: Subramonian, Arjun, et autres
Publié: (2024)

Side-by-side Comparison Amplifies Dialect Bias in Language Models
par: Kondapally, Kritee, et autres
Publié: (2026)

Generative Explanations for Program Synthesizers
par: Nazari, Amirmohammad, et autres
Publié: (2024)

Repetition Improves Language Model Embeddings
par: Springer, Jacob Mitchell, et autres
Publié: (2024)

Still Not Quite There! Evaluating Large Language Models for Comorbid Mental Health Diagnosis
par: Hengle, Amey, et autres
Publié: (2024)

IYKYK (But AI Doesn't): Automated Content Moderation Does Not Capture Communities' Heterogeneous Attitudes Towards Reclaimed Language
par: Chance, Christina, et autres
Publié: (2026)

OATH-Frames: Characterizing Online Attitudes Towards Homelessness with LLM Assistants
par: Ranjit, Jaspreet, et autres
Publié: (2024)

Every Language Model Has a Forgery-Resistant Signature
par: Finlayson, Matthew, et autres
Publié: (2025)

ChEmREF: Evaluating Language Model Readiness for Chemical Emergency Response
par: Surana, Risha, et autres
Publié: (2025)

Agree to Disagree? A Meta-Evaluation of LLM Misgendering
par: Subramonian, Arjun, et autres
Publié: (2025)

Unveiling Biases while Embracing Sustainability: Assessing the Dual Challenges of Automatic Speech Recognition Systems
par: Kulkarni, Ajinkya, et autres
Publié: (2025)

Why Fine-Tuning Encourages Hallucinations and How to Fix It
par: Kaplan, Guy, et autres
Publié: (2026)

Are We Automating the Joy Out of Work? Designing AI to Augment Work, Not Meaning
par: Ranjit, Jaspreet, et autres
Publié: (2026)

Music for All: Representational Bias and Cross-Cultural Adaptability of Music Generation Models
par: Mehta, Atharva, et autres
Publié: (2025)

Overtrained Language Models Are Harder to Fine-Tune
par: Springer, Jacob Mitchell, et autres
Publié: (2025)

Understanding the Influence of Synthetic Data for Text Embedders
par: Springer, Jacob Mitchell, et autres
Publié: (2025)

Do Compact SSL Backbones Matter for Audio Deepfake Detection? A Controlled Study with RAPTOR
par: Kulkarni, Ajinkya, et autres
Publié: (2026)

Annotations Mitigate Post-Training Mode Collapse
par: Springer, Jacob Mitchell, et autres
Publié: (2026)

Uncovering Gaps in How Humans and LLMs Interpret Subjective Language
par: Jones, Erik, et autres
Publié: (2025)

Sharpness-Aware Pretraining Mitigates Catastrophic Forgetting
par: Watts, Ishaan, et autres
Publié: (2026)

Extend Model Merging from Fine-Tuned to Pre-Trained Large Language Models via Weight Disentanglement
par: Yu, Le, et autres
Publié: (2024)