Copertina

Negation Neglect: When models fail to learn negations in training

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Mayne, Harry, McKinney, Lev, Dubiński, Jan, Karvonen, Adam, Chua, James, Evans, Owain
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Computation and Language Artificial Intelligence Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
di: Berglund, Lukas, et al.
Pubblicazione: (2023)

Thought Crime: Backdoors and Emergent Misalignment in Reasoning Models
di: Chua, James, et al.
Pubblicazione: (2025)

Tell me about yourself: LLMs are aware of their learned behaviors
di: Betley, Jan, et al.
Pubblicazione: (2025)

Can sparse autoencoders be used to decompose and interpret steering vectors?
di: Mayne, Harry, et al.
Pubblicazione: (2024)

Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers
di: Karvonen, Adam, et al.
Pubblicazione: (2025)

Lessons from Studying Two-Hop Latent Reasoning
di: Balesni, Mikita, et al.
Pubblicazione: (2024)

Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
di: Dubiński, Jan, et al.
Pubblicazione: (2026)

Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs
di: Betley, Jan, et al.
Pubblicazione: (2025)

School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs
di: Taylor, Mia, et al.
Pubblicazione: (2025)

Robustly Improving LLM Fairness in Realistic Settings via Interpretability
di: Karvonen, Adam, et al.
Pubblicazione: (2025)

Looking Inward: Language Models Can Learn About Themselves by Introspection
di: Binder, Felix J, et al.
Pubblicazione: (2024)

Connecting the Dots: LLMs can Infer and Verbalize Latent Structure from Disparate Training Data
di: Treutlein, Johannes, et al.
Pubblicazione: (2024)

The Consciousness Cluster: Emergent preferences of Models that Claim to be Conscious
di: Chua, James, et al.
Pubblicazione: (2026)

What can large language models do for sustainable food?
di: Thomas, Anna T., et al.
Pubblicazione: (2025)

LLMs Don't Know Their Own Decision Boundaries: The Unreliability of Self-Generated Counterfactual Explanations
di: Mayne, Harry, et al.
Pubblicazione: (2025)

Subliminal Learning: Language models transmit behavioral traits via hidden signals in data
di: Cloud, Alex, et al.
Pubblicazione: (2025)

LINGOLY-TOO: Disentangling Reasoning from Knowledge with Templatised Orthographic Obfuscation
di: Khouja, Jude, et al.
Pubblicazione: (2025)

When to Memorize and When to Stop: Gated Recurrent Memory for Long-Context Reasoning
di: Sheng, Leheng, et al.
Pubblicazione: (2026)

Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs
di: Laine, Rudolf, et al.
Pubblicazione: (2024)

Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
di: Casademunt, Helena, et al.
Pubblicazione: (2025)

Towards LLM-based optimization compilers. Can LLMs learn how to apply a single peephole optimization? Reasoning is all LLMs need!
di: Fang, Xiangxin, et al.
Pubblicazione: (2024)

Why does in-context learning fail sometimes? Evaluating in-context learning on open and closed questions
di: Li, Xiang, et al.
Pubblicazione: (2024)

Efficient LLM Moderation with Multi-Layer Latent Prototypes
di: Chrabąszcz, Maciej, et al.
Pubblicazione: (2025)

An Independent Safety Evaluation of Kimi K2.5
di: Yong, Zheng-Xin, et al.
Pubblicazione: (2026)

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
di: Betley, Jan, et al.
Pubblicazione: (2025)

From Fallback to Frontline: When Can LLMs be Superior Annotators of Human Perspectives?
di: Amin, Hasan, et al.
Pubblicazione: (2026)

Negating Negatives: Alignment with Human Negative Samples via Distributional Dispreference Optimization
di: Duan, Shitong, et al.
Pubblicazione: (2024)

CoCoNUTS: Concentrating on Content while Neglecting Uninformative Textual Styles for AI-Generated Peer Review Detection
di: Chen, Yihan, et al.
Pubblicazione: (2025)

Can Language Models Explain Their Own Classification Behavior?
di: Sherburn, Dane, et al.
Pubblicazione: (2024)

Response-free item difficulty modelling for multiple-choice items with fine-tuned transformers: Component-wise representation and multi-task learning
di: Netík, Jan, et al.
Pubblicazione: (2026)

Do Language Models Know When They're Hallucinating References?
di: Agrawal, Ayush, et al.
Pubblicazione: (2023)

Concept Influence: Leveraging Interpretability to Improve Performance and Efficiency in Training Data Attribution
di: Kowal, Matthew, et al.
Pubblicazione: (2026)

Empaths at SemEval-2025 Task 11: Retrieval-Augmented Approach to Perceived Emotions Prediction
di: Morozov, Lev, et al.
Pubblicazione: (2025)

When Context Leads but Parametric Memory Follows in Large Language Models
di: Tao, Yufei, et al.
Pubblicazione: (2024)

Theory of Mind and Self-Attributions of Mentality are Dissociable in LLMs
di: Kim, Junsol, et al.
Pubblicazione: (2026)

When a language model is optimized for reasoning, does it still show embers of autoregression? An analysis of OpenAI o1
di: McCoy, R. Thomas, et al.
Pubblicazione: (2024)

Retracing the Past: LLMs Emit Training Data When They Get Lost
di: Ko, Myeongseob, et al.
Pubblicazione: (2025)

MinorBench: A hand-built benchmark for content-based risks for children
di: Khoo, Shaun, et al.
Pubblicazione: (2025)

Machine learning methods fail to provide cohesive atheoretical construction of personality traits from semantic embeddings
di: Bouguettaya, Ayoub, et al.
Pubblicazione: (2025)

ChildEval: When large language models meet children's personalities
di: Luo, Yanyan, et al.
Pubblicazione: (2026)