Related papers: Direct Evidence Delay with A Task Decreases Workin…
We discuss the two moments of human cognition, namely, apprehension (A), whereby a coherent perception emerges from the recruitment of neuronal groups, and judgment(B),that entails the comparison of two apprehensions acquired at different…
Working becomes harder as we grow tired or bored. I model individuals who underestimate these changes in marginal disutility -- as implied by "projection bias" -- when deciding whether or not to continue working. This bias causes people's…
It is often difficult to hand-specify what the correct reward function is for a task, so researchers have instead aimed to learn reward functions from human behavior or feedback. The types of behavior interpreted as evidence of the reward…
Pretraining language models directly on web-scale corpora is the de facto paradigm. We study an alternative where the model is initially exposed to abstract structured data to ease the subsequent acquisition of rich semantic knowledge, much…
Proof Blocks is a software tool that provides students with a scaffolded proof-writing experience, allowing them to drag and drop prewritten proof lines into the correct order instead of starting from scratch. In this paper we describe a…
When we encounter a new person or place, we may easily encode it into our memories, or we may quickly forget it. Recent work finds that this likelihood of encoding a given entity - memorability - is highly consistent across viewers and…
In reinforcement learning, a decision needs to be made at some point as to whether it is worthwhile to carry on with the learning process or to terminate it. In many such situations, stochastic elements are often present which govern the…
This article presents results from an experiment in which 44 human subjects interact with a dynamic system to perform 40 trials of a command-following task. The reference command is unpredictable and different on each trial, but all…
Reasoning is a core capability of large language models, yet how multi-step reasoning is learned and executed remains unclear. We study this question in a controlled cellular-automata (1dCA) framework that excludes memorisation by using…
Language models generate responses by producing a series of tokens in immediate succession: the $(K+1)^{th}$ token is an outcome of manipulating $K$ hidden vectors per layer, one vector per preceding token. What if instead we were to let…
Machine unlearning has emerged as a key component in ensuring ``Right to be Forgotten'', enabling the removal of specific data points from trained models. However, even when the unlearning is performed without poisoning the forget-set…
We modify a canonical experimental design to identify the effectiveness of retractions. Comparing beliefs after retractions to beliefs (a) without the retracted information and (b) after equivalent new information, we find that retractions…
Experience replay is a key technique behind many recent advances in deep reinforcement learning. Allowing the agent to learn from earlier memories can speed up learning and break undesirable temporal correlations. Despite its wide-spread…
This work introduces a new task preemption primitive for Hadoop, that allows tasks to be suspended and resumed exploiting existing memory management mechanisms readily available in modern operating systems. Our technique fills the gap that…
Language models are increasingly being deployed as user simulators, but their memory is far more reliable than that of real users. To measure this gap, we run a series of classic memory experiments from psychology on both humans and…
In this work, we address the symptoms of cognitive depletion as they relate to generalized knowledge workers. We unify previous findings within a single analytical model of cognitive depletion. Our purpose is to develop a model that will…
Backtesting large language models on historical events requires reasoning exclusively from information available before a specified cutoff date. Yet models routinely leak post-cutoff knowledge from pre-training into their reasoning,…
Leveraging contextual knowledge has become standard practice in automated claim verification, yet the impact of temporal reasoning has been largely overlooked. Our study demonstrates that time positively influences the claim verification…
The impressive capability of modern text-to-image models to generate realistic visuals has come with a serious drawback: they can be misused to create harmful, deceptive or unlawful content. This has accelerated the push for machine…
Aside from crawling, indexing, and querying RDF data centrally, Linked Data principles allow for processing SPARQL queries on-the-fly by dereferencing URIs. Proposed link-traversal query approaches for Linked Data have the benefits of…