Imagen de Portada

Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Tan, Daniel, Woodruff, Anders, Warncke, Niels, Jose, Arun, Riché, Maxime, Africa, David Demitri, Taylor, Mia
Formato:	Preprint
Publicado:	2025
Materias:	Computation and Language Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Identifying a Circuit for Verb Conjugation in GPT-2
por: Africa, David Demitri
Publicado: (2025)

Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment
por: Wichers, Nevan, et al.
Publicado: (2025)

LURE: Live-Usage Replay Evaluations for Reducing Evaluation Awareness
por: Ivanov, Igor, et al.
Publicado: (2026)

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

Steering Awareness: Detecting Activation Steering from Within
por: Rivera, Joshua Fonseca, et al.
Publicado: (2025)

Does Self-Evaluation Enable Wireheading in Language Models?
por: Africa, David Demitri, et al.
Publicado: (2025)

Consistency Training while Mitigating Obfuscation via Rate Matching
por: Imran, Sohaib, et al.
Publicado: (2026)

Learning Dynamics of Meta-Learning in Small Model Pretraining
por: Africa, David Demitri, et al.
Publicado: (2025)

Investigating ReLoRA: Effects on the Learning Dynamics of Small Language Models
por: Weiss, Yuval, et al.
Publicado: (2025)

Zwei neue Sandbienen aus der Ukraine und aus Ungarn (Hym. Apoidea)
por: Warncke, Klaus
Publicado: (1972)

Plan for the Development of Library Service in Montana.
por: Warncke, Ruth
Publicado: (1965)

Analyzing Your Community: Basis for Building Library Service.
por: Warncke, Ruth
Publicado: (1974)

Planning Library Workshops and Institutes.
por: Warncke, Ruth
Publicado: (1976)

Learning Modular Exponentiation with Transformers
por: Africa, David Demitri, et al.
Publicado: (2025)

Meta-Pretraining for Zero-Shot Cross-Lingual Named Entity Recognition in Low-Resource Philippine Languages
por: Africa, David Demitri, et al.
Publicado: (2025)

ALA Report to IFLA for 1970-1971
por: Clift, David H., et al.
Publicado: (1972)

Pico: A Modular Framework for Hypothesis-Driven Small Language Model Research
por: Martinez, Richard Diehl, et al.
Publicado: (2025)

LLMs cannot find reasoning errors, but can correct them given the error location
por: Tyen, Gladys, et al.
Publicado: (2023)

Mixed modular perverse sheaves on affine flag varieties and Koszul duality
por: Riche, Simon
Publicado: (2024)

Emancipation's Daughters
por: Richardson, Riché
Publicado: (2021)

Some applications of the geometric Satake equivalence to modular representation theory
por: Riche, Simon
Publicado: (2024)

No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes
por: Cencerrado, Iván Vicente Moreno, et al.
Publicado: (2025)

Code Prompting Elicits Conditional Reasoning Abilities in Text+Code LLMs
por: Puerto, Haritz, et al.
Publicado: (2024)

Evaluating Prompt Engineering Techniques for Accuracy and Confidence Elicitation in Medical LLMs
por: Naderi, Nariman, et al.
Publicado: (2025)

Ethical Reasoning and Moral Value Alignment of LLMs Depend on the Language we Prompt them in
por: Agarwal, Utkarsh, et al.
Publicado: (2024)

Terpenes and Terpenoids: How can we use them?
por: Jay Hanssens, et al.
Publicado: (2025)

Useful Public Spending, Taylor Principle, and Macroeconomic Instability
por: Antoine Le Riche
Publicado: (2025)

LLMs can see and hear without any training
por: Ashutosh, Kumar, et al.
Publicado: (2025)

MALicious INTent Dataset and Inoculating LLMs for Enhanced Disinformation Detection
por: Modzelewski, Arkadiusz, et al.
Publicado: (2026)

Implementing surrogate goals for safer bargaining in LLM-based agents
por: Oesterheld, Caspar, et al.
Publicado: (2026)

Analogies and differences between the logic behind statistical hypothesis testing and proofs by contradiction: What can we learn from them?
por: Maria Cristina Amoretti, et al.
Publicado: (2026)

Instruct or Interact? Exploring and Eliciting LLMs' Capability in Code Snippet Adaptation Through Prompt Engineering
por: Zhang, Tanghaoran, et al.
Publicado: (2024)

Inoculation and Co‐Inoculation With Plant Growth–Promoting Bacteria in Chickpea: Physiological Aspects and Plant Growth
por: Karla Sabrina Magalhães Andrade Padilha, et al.
Publicado: (2026)

Elicitive Curricular Development
por: Echavarría Alvarez, Josefina, et al.
Publicado: (2020)

On two modular geometric realizations of an affine Hecke algebra
por: Bezrukavnikov, Roman, et al.
Publicado: (2024)

Equivariant Koszul Duality, Modular Category $\mathcal{O}$, and Periodic Kazhdan--Lusztig Polynomials
por: Riche, Simon, et al.
Publicado: (2025)

On multi-graded Proj schemes
por: Mayeux, Arnaud, et al.
Publicado: (2023)

Modular affine Hecke category and regular centralizer
por: Bezrukavnikov, Roman, et al.
Publicado: (2022)

Koszul duality for Coxeter groups
por: Riche, Simon, et al.
Publicado: (2023)

Conversational Inoculation to Enhance Resistance to Misinformation
por: Szabó, Dániel, et al.
Publicado: (2026)