Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers

Y Jiang, G Rajendran, P Ravikumar… - Advances in Neural …, 2025 - proceedings.neurips.cc
Abstract Large Language Models (LLMs) have the capacity to store and recall facts. Through
experimentation with open-source models, we observe that this ability to retrieve facts can …

Chaos with keywords: exposing large language models sycophancy to misleading keywords and evaluating defense strategies

A Rrv, N Tyagi, MN Uddin, N Varshney… - Findings of the …, 2024 - aclanthology.org
This study explores the sycophantic tendencies of Large Language Models (LLMs), where
these models tend to provide answers that match what users want to hear, even if they are …

Chaos with Keywords: Exposing Large Language Models Sycophantic Hallucination to Misleading Keywords and Evaluating Defense Strategies

A RRV, N Tyagi, MN Uddin, N Varshney… - arxiv preprint arxiv …, 2024 - arxiv.org
This study explores the sycophantic tendencies of Large Language Models (LLMs), where
these models tend to provide answers that match what users want to hear, even if they are …

Large Language Models are In-context Teachers for Knowledge Reasoning

J Zhao, Z Yao, Z Yang, H Yu - arxiv preprint arxiv:2311.06985, 2023 - arxiv.org
In this work, we study in-context teaching (ICT), where a teacher provides in-context
example rationales to teach a student to reason over unseen cases. Human teachers are …

Associative memory inspires improvements for in-context learning using a novel attention residual stream architecture

TF Burns, T Fukai, CJ Earls - arxiv preprint arxiv:2412.15113, 2024 - arxiv.org
Large language models (LLMs) demonstrate an impressive ability to utilise information
within the context of their input sequences to appropriately respond to data unseen by the …