Copertina

Semantic Gravity Wells: Why Negative Constraints Backfire

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autore principale:	Rana, Shailesh
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Artificial Intelligence Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The Shape of Wisdom: Decision Trajectories in Language Models
di: Rana, Shailesh
Pubblicazione: (2026)

When Chain-of-Thought Backfires: Evaluating Prompt Sensitivity in Medical Language Models
di: Sadanandan, Binesh, et al.
Pubblicazione: (2026)

Emulated Disalignment: Safety Alignment for Large Language Models May Backfire!
di: Zhou, Zhanhui, et al.
Pubblicazione: (2024)

NCO: A Versatile Plug-in for Handling Negative Constraints in Decoding
di: Jin, Hyundong, et al.
Pubblicazione: (2026)

Negation Triplet Extraction with Syntactic Dependency and Semantic Consistency
di: Shi, Yuchen, et al.
Pubblicazione: (2024)

Enhancing Semantics in Multimodal Chain of Thought via Soft Negative Sampling
di: Zheng, Guangmin, et al.
Pubblicazione: (2024)

Semantic Adapter for Universal Text Embeddings: Diagnosing and Mitigating Negation Blindness to Enhance Universality
di: Cao, Hongliu
Pubblicazione: (2025)

Mitigating Semantic Leakage in Cross-lingual Embeddings via Orthogonality Constraint
di: Ki, Dayeon, et al.
Pubblicazione: (2024)

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

When Incentives Backfire, Data Stops Being Human
di: Santy, Sebastin, et al.
Pubblicazione: (2025)

Alignment Backfire: Language-Dependent Reversal of Safety Interventions Across 16 Languages in LLM Multi-Agent Systems
di: Fukui, Hiroki
Pubblicazione: (2026)

Which bird does not have wings: Negative-constrained KGQA with Schema-guided Semantic Matching and Self-directed Refinement
di: Shim, Midan, et al.
Pubblicazione: (2026)

Under Pressure: Emotional Framing Induces Measurable Behavioral Shifts and Structured Internal Geometry in Small Language Models
di: Usman, Rana Muhammad
Pubblicazione: (2026)

Semantic Integrity Constraints: Declarative Guardrails for AI-Augmented Data Processing Systems
di: Lee, Alexander W., et al.
Pubblicazione: (2025)

Correcting Negative Bias in Large Language Models through Negative Attention Score Alignment
di: Yu, Sangwon, et al.
Pubblicazione: (2024)

Semantic Anchors in In-Context Learning: Why Small LLMs Cannot Flip Their Labels
di: Kumar, Anantha Padmanaban Krishna
Pubblicazione: (2025)

A Pseudo-Semantic Loss for Autoregressive Models with Logical Constraints
di: Ahmed, Kareem, et al.
Pubblicazione: (2023)

WellDunn: On the Robustness and Explainability of Language Models and Large Language Models in Identifying Wellness Dimensions
di: Mohammadi, Seyedali, et al.
Pubblicazione: (2024)

Why are LLMs' abilities emergent?
di: Havlík, Vladimír
Pubblicazione: (2025)

Why Attend to Everything? Focus is the Key
di: Yao, Hengshuai, et al.
Pubblicazione: (2026)

Generating Diverse Negations from Affirmative Sentences
di: Vasquez, Darian Rodriguez, et al.
Pubblicazione: (2024)

AraTable: Benchmarking LLMs' Reasoning and Understanding of Arabic Tabular Data
di: Alshaikh, Rana, et al.
Pubblicazione: (2025)

Fake Alignment: Are LLMs Really Aligned Well?
di: Wang, Yixu, et al.
Pubblicazione: (2023)

Reasoning Models Reason Well, Until They Don't
di: Rameshkumar, Revanth, et al.
Pubblicazione: (2025)

Why Slop Matters
di: Kommers, Cody, et al.
Pubblicazione: (2025)

Large-Scale Constraint Generation -- Can LLMs Parse Hundreds of Constraints?
di: Boffa, Matteo, et al.
Pubblicazione: (2025)

The Impact of Negated Text on Hallucination with Large Language Models
di: Seo, Jaehyung, et al.
Pubblicazione: (2025)

How Well Do LLMs Understand Tunisian Arabic?
di: Mahdi, Mohamed
Pubblicazione: (2025)

Why Chain of Thought Fails in Clinical Text Understanding
di: Wu, Jiageng, et al.
Pubblicazione: (2025)

Why is constrained neural language generation particularly challenging?
di: Garbacea, Cristina, et al.
Pubblicazione: (2022)

Why Braking? Scenario Extraction and Reasoning Utilizing LLM
di: Wu, Yin, et al.
Pubblicazione: (2025)

Carrot and Stick: Inducing Self-Motivation with Positive & Negative Feedback
di: Sohn, Jimin, et al.
Pubblicazione: (2024)

Mitigating the Negative Impact of Over-association for Conversational Query Production
di: Wang, Ante, et al.
Pubblicazione: (2024)

Teaching with Lies: Curriculum DPO on Synthetic Negatives for Hallucination Detection
di: Pandit, Shrey, et al.
Pubblicazione: (2025)

MCJudgeBench: A Benchmark for Constraint-Level Judge Evaluation in Multi-Constraint Instruction Following
di: Lee, Jaeyun, et al.
Pubblicazione: (2026)

How Well Do Large Language Models Truly Ground?
di: Lee, Hyunji, et al.
Pubblicazione: (2023)

Why Retrieval-Augmented Generation Fails: A Graph Perspective
di: Guo, Kai, et al.
Pubblicazione: (2026)

Surgical Feature-Space Decomposition of LLMs: Why, When and How?
di: Chavan, Arnav, et al.
Pubblicazione: (2024)

Look Within, Why LLMs Hallucinate: A Causal Perspective
di: Li, He, et al.
Pubblicazione: (2024)

Towards Minimal Targeted Updates of Language Models with Targeted Negative Training
di: Zhang, Lily H., et al.
Pubblicazione: (2024)