LESSWRONGTags
LW

Eliciting Latent Knowledge (ELK)

•

Applied to [Paper Blogpost] When Your AIs Deceive You: Challenges with Partial Observability in RLHF by Leon Lang 1mo ago

•

Applied to Clarifying Alignment Fundamentals Through the Lens of Ontology by eternal/ephemera 1mo ago

•

Applied to Mechanistic Anomaly Detection Research Update by Nora Belrose 3mo ago

•

Applied to Covert Malicious Finetuning by Tony Wang 5mo ago

•

Applied to "What the hell is a representation, anyway?" | Clarifying AI interpretability with tools from philosophy of cognitive science | Part 1: Vehicles vs. contents by IwanWilliams 5mo ago

•

Applied to Is This Lie Detector Really Just a Lie Detector? An Investigation of LLM Probe Specificity. by Josh Levy 6mo ago

•

Applied to CCS on compound sentences by Artyom Karpov 7mo ago

•

Applied to AXRP Episode 29 - Science of Deep Learning with Vikrant Varma by DanielFilan 7mo ago

•

Applied to Finding the estimate of the value of a state in RL agents by Clément Dumas 7mo ago

•

Applied to Auditing LMs with counterfactual search: a tool for control and ELK by Jacob Pfau 9mo ago

•

Applied to Striking Implications for Learning Theory, Interpretability — and Safety? by RogerDearnaley 11mo ago

•

Applied to Measurement tampering detection as a special case of weak-to-strong generalization by ryan_greenblatt 11mo ago

•

Applied to Discussion: Challenges with Unsupervised LLM Knowledge Discovery by Seb Farquhar 1y ago

•

Applied to Betting on what is un-falsifiable and un-verifiable by Abhimanyu Pallavi Sudhir 1y ago

•

Applied to Eliciting Latent Knowledge in Comprehensive AI Services Models by acabodi 1y ago

•

Applied to Robustness of Contrast-Consistent Search to Adversarial Prompting by Nandi 1y ago

•

Applied to Discovering Latent Knowledge in the Human Brain: Part 1 – Clarifying the concepts of belief and knowledge by Joseph Emerson 1y ago

•

Applied to Attributing to interactions with GCPD and GWPD by jenny 1y ago

•

Applied to A personal explanation of ELK concept and task. by Zeyu Qin 1y ago