Image de couverture de livre

Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Wichers, Nevan, Ebtekar, Aram, Azarbal, Ariana, Gillioz, Victor, Ye, Christine, Ryd, Emil, Rathi, Neil, Sleight, Henry, Mallen, Alex, Roger, Fabien, Marks, Samuel
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Recontextualization Mitigates Specification Gaming without Modifying the Specification
par: Azarbal, Ariana, et autres
Publié: (2025)

Model Spec Midtraining: Improving How Alignment Training Generalizes
par: Li, Chloe, et autres
Publié: (2026)

Gradient-Based Language Model Red Teaming
par: Wichers, Nevan, et autres
Publié: (2024)

Foundations of algorithmic thermodynamics
par: Ebtekar, Aram, et autres
Publié: (2023)

Golden Handcuffs make safer AI agents
par: Ebtekar, Aram, et autres
Publié: (2026)

Visualizing Neural Network Imagination
par: Wichers, Nevan, et autres
Publié: (2024)

Toward Universal Laws of Outlier Propagation
par: Ebtekar, Aram, et autres
Publié: (2025)

Fine Flood Forecasts: Incorporating local data into global models through fine-tuning
par: Ryd, Emil, et autres
Publié: (2025)

Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time
par: Tan, Daniel, et autres
Publié: (2025)

The momentum-space conformal bootstrap in 2d
par: Gillioz, Marc
Publié: (2025)

Dynamic Relation Inference via Verb Embeddings
par: Suissa, Omri, et autres
Publié: (2025)

Eliciting Secret Knowledge from Language Models
par: Cywiński, Bartosz, et autres
Publié: (2025)

Beyond Sparse Rewards: Enhancing Reinforcement Learning with Language Model Critique in Text Generation
par: Cao, Meng, et autres
Publié: (2024)

Fusion-Eval: Integrating Assistant Evaluators with LLMs
par: Shu, Lei, et autres
Publié: (2023)

Towards eliciting latent knowledge from LLMs with mechanistic interpretability
par: Cywiński, Bartosz, et autres
Publié: (2025)

All Code, No Thought: Current Language Models Struggle to Reason in Ciphered Language
par: Guo, Shiyuan, et autres
Publié: (2025)

Analysis of Parallel Boarding Methods in a Multi-Aisle Flying Wing Aircraft
par: Ryd, Emil, et autres
Publié: (2024)

Intensive language contact in the Caucasus
par: Wichers Schreur, Jesse
Publié: (2026)

Problems of Associations Accrediting Preparation Programs for School Media Professionals.
par: Wichers, Jean Elaine
Publié: (1982)

Heart of the Humanities Program
par: Wichers, Jean Elaine
Publié: (1976)

Removing Sandbagging in LLMs by Training with Weak Supervision
par: Ryd, Emil, et autres
Publié: (2026)

El ascenso de la economía del comportamiento: misbehaving, de Richard Thaler*
par: Cass R. Sunsteitf
Publié: (2019)

Shaping capabilities with token-level data filtering
par: Rathi, Neil, et autres
Publié: (2026)

A large synthetic dataset for machine learning applications in power transmission grids
par: Gillioz, Marc, et autres
Publié: (2024)

IPAdapter-Instruct: Resolving Ambiguity in Image-based Conditioning using Instruct Prompts
par: Rowles, Ciara, et autres
Publié: (2024)

Los bosques en 50 años : de los planes de ordenación a la silvicultura electrónica (Primera parte)
par: Mallén, Carlos
Publié: (2002)

Los bosques en 50 años : de la recolecta al cultivo, plantaciones comerciales y no comerciales (segunda parte)
par: Mallén, Carlos
Publié: (2002)

Real time anomalies detection on video
par: Poirier, Fabien
Publié: (2024)

PROTECT: Protein circadian time prediction using unsupervised learning
par: Ogholbake, Aram Ansary, et autres
Publié: (2025)

VIP: Vision Instructed Pre-training for Robotic Manipulation
par: Li, Zhuoling, et autres
Publié: (2024)

Prompt-aligned Gradient for Prompt Tuning
par: Zhu, Beier, et autres
Publié: (2022)

México en la evaluación de los bosques del mundo
par: Mallén Rivera, Carlos
Publié: (2003)

Desarrollo histórico del pensamiento político / Rubén Salazar Mallén
par: Salazar Mallén, Rubén

Conservar la biodiversidad, ¿mito o realidad? / Isabel Ruiz Mallén
par: Ruiz Mallén, Isabel

Año 2002 montañas, cumbres, leyes y religión
par: Mallén Rivera, Carlos
Publié: (2003)

El Hegel de Hegel y el Hegel de Marx / Rubén Salazar Mallén
par: Salazar Mallén, Rubén

La polémica chinosoviética y la revolución proletaria / Rubén Salazar Mallén
par: Salazar Mallén, Rubén
Publié: (1965)

La orientación inegable de la biotecnología : de las estacas a los organismos genéticamente transformados
par: Mallén Rivera, Carlos
Publié: (2002)

La orientación inegable de la biotecnología (De las estacas a los organismos genéticamente transformados (2a. parte)
par: Mallén Rivera, Carlos
Publié: (2002)

Virus computacionales : Un enfoque objetivo / Guillermo M. Mallen
par: Mallen, Guillermo M