Mechanisms
Taxonomy
Memory Mechanisms Taxonomy
29 constructs of human memory, each empirically established in humans and implementable in current AI memory architectures, grouped by the purpose each one serves.
The structure
Select a group to open the level beneath it; a mechanism opens its own page.
How were memory constructs selected?
Inclusion Criteria
Using Tulving's (1972) episodic/semantic distinction and Squire's (2004) declarative/non-declarative taxonomy as a starting point, the memory literature on systems and processes was surveyed and each candidate was analyzed against two criteria.
-
01
Empirically established in humans, evidenced by a stable definition with agreed-upon examples, an identified neural substrate or lesion dissociation, and quantitative or meta-analytic effect sizes.
-
02
Implementable and measurable in current AI memory architectures, whether or not it has yet been measured.
How the candidates sorted against the two criteria is shown on the Method page.
What was included and excluded?
Included Constructs
29 constructs were identified across the human memory literature meeting both criteria above.
Constructs were grouped by what kind of purpose each one is serving rather than by the declarative and non-declarative division.
- Memory Systems
- 8 Memory Systems Retains Squire's (2004) organization by kind of representation.
- Representational Structures
- 4 Representational Structures Organizing schemes imposed over stored content.
- Memory Processes
- 12 Memory Processes Operations performed on stored content.
- Metacognition
- 5 Metacognition Second-order judgments about stored content.
Excluded Constructs
Examples of candidates that were considered and excluded, either for failing one of the two criteria or for being the wrong unit of comparison. They are listed because their exclusion reflects the boundaries of the criteria rather than the limits of the review.
Sensory Memory Fails criterion 2
Sensory memory meets the first criterion, with an established paradigm (Sperling, 1960), documented decay parameters, and reported effect sizes, but fails the second. Its counterpart in an AI system is input tokenization, which is not memory in any sense examined here. The exclusion is noted explicitly because existing surveys of AI memory routinely include sensory memory as a fifth system, on the strength of that mapping.
Encoding Wrong unit: a stage, not a function
Encoding was excluded as a processing stage rather than a function. Stages describe when an operation occurs rather than what it accomplishes, which makes them the wrong unit for comparison across substrates. Encoding specificity is treated instead as a cross-cutting factor.
Transactive Memory Wrong unit: a group, not a person
Transactive memory was excluded as a different unit of analysis rather than a different construct. Wegner's account describes encoding, storage, and retrieval distributed across members of a group, with individuals holding a directory of who knows what (Wegner et al., 1991). Because the system is the group rather than the person, it carries no neural substrate of the kind the first criterion requires, and it is not a factor modulating individual memory processes in the way the cross-cutting factors are.
It is retained as a framing device for the use-case comparison, where the relevant question is not what memory functions an AI system should have but what a transactive partner requires in order to be a good one.
Cross-Cutting Factors
Cross-Cutting Factors
A third category holds influences that shape multiple memory constructs without being constructs themselves.
They are not functions the memory system performs but conditions under which it performs them. A factor was included where it meets all three of the following:
-
01
It has documented effects on two or more of the constructs.
-
02
It carries quantitative or meta-analytic evidence for those effects.
-
03
It raises a question about AI memory that the construct-level analysis cannot address.
Some factors have a well-localized neural substrate, as emotional modulation does in the basolateral amygdala (McGaugh, 2004); others, such as cultural variability, have none. What distinguishes a factor from a construct here is not anatomy but scope: a factor conditions how multiple constructs operate rather than being an operation itself.
Examples
Cultural Variability
Human memory is shaped by culture, which influences the age of earliest recollection and the content and style of autobiographical memories (Wang, 2001), is transmitted developmentally through the ways caregivers reminisce with children about the past (Nelson & Fivush, 2004), and appears to moderate whether detailed episodic recall supports or undermines well-being (Wang et al., 2018; see Wang, 2021, for review). Cultural variation is not a property of one memory mechanism but a factor operating across them, from perceptual encoding through autobiographical narrative.
The AI question it raises
Large language models are trained on text scraped from the web, which overrepresents dominant voices rather than representative ones: GPT-2 was trained on web pages linked from Reddit, whose US users are 67% male and 64% aged 18 to 29, and only 8.8 to 15% of Wikipedia editors are women (Bender et al., 2021). Filtering compounds the skew rather than correcting it, since blocklist-based cleaning removes 42% of documents in African American English and 32% in Hispanic-aligned English, against 6.2% of White-aligned English, leaving 97.8% of the resulting corpus in the latter (Dodge et al., 2021). What is absorbed is not a separable contaminant but the meaning of the language itself: word embeddings trained on ordinary web text reproduce the full range of human associations, from the morally neutral to the prejudicial, at effect sizes matching those measured in people (Caliskan et al., 2017). The consequences are measurable in current systems. Models produce outputs resembling the cultural values of English-speaking and Protestant European countries when prompted without a specified cultural identity (Tao et al., 2024); their expressed opinions diverge from those of the US public by roughly the margin separating Democrats and Republicans on climate change, with older, widowed, and religiously observant respondents least well represented (Santurkar et al., 2023); and they hold covert stereotypes about speakers of African American English more negative than any human stereotypes about African Americans ever experimentally recorded (Hofmann et al., 2024).
Emotional Modulation
Emotional arousal modulates how strongly an event is retained, acting on storage systems already in use rather than through a separate emotional memory store (McGaugh, 2004). The mechanism is contested: on an alternative account the amygdala binds items to emotional responses, and those bindings are simply forgotten more slowly than the item-context bindings the hippocampus supports (Yonelinas & Ritchey, 2015). The effect is selective rather than uniform, but the selection is by priority rather than spatial position: arousal amplifies whatever already dominates the competition for processing and suppresses what does not, producing a winner-take-more pattern in both perception and consolidation (Mather & Sutherland, 2011). Because arousal acts at encoding and after it (McGaugh, 2004), and acute stress separately impairs retrieval, more so for emotional than neutral material (Shields et al., 2017), it cuts across four rows rather than sitting in one. Retrieval is also a starting point and not only an end point, since recalling an emotional event reshapes the trace available to later recall (Kensinger & Ford, 2020), which is consistent with Squire's (2004) treatment of arousal modulation as affecting both declarative and non-declarative memory. Clinically, the same modulation is implicated in the involuntary character of trauma memory (Brewin, 2014).
The AI question it raises
AI memory systems have no analogue to arousal modulation. When a system stores something a user has said, it later finds that material by how closely a numerical representation of the current question matches those of what was stored, a similarity calculation with no reference to how the material was acquired. Agent architectures add two adjustments to this: a recency weight that fades with time, and an importance rating from 1 to 10 that a language model assigns at the moment of storage by being asked how significant the item seems (Park et al., 2023). That rating is the closest counterpart to emotional arousal, and it does more than order search results, since accumulated importance also determines when the system pauses to draw higher-level conclusions from recent material. But it is a label attached to a stored item, not a state of the system that changes how storage happens. Emotional content in what a user writes does change how these systems behave, both on psychological self-report measures (Ben-Zion et al., 2025; Coda-Forno et al., 2023) and on task performance (Li et al., 2023), yet it does so by shaping the response being generated in the moment rather than by altering what is retained. The user's emotional state therefore has no privileged route into what is stored or what later becomes accessible.
Encoding Specificity
What is retained depends on how material is processed at input, not merely on whether it is attended. Semantic or elaborative processing produces substantially better retention than perceptual processing of the same material, with recognition varying between .14 and .96 as a function of the orienting question alone (Craik & Tulving, 1975). The effect is not reducible to time on task: a structural task that took twice as long as a semantic one still yielded worse recognition, and the advantage persists under intentional learning and when participants are paid more to remember shallowly encoded items. Attention is not a separate constraint so much as a determinant of the same variable, since dividing attention at encoding produces shallower encoding; encoding is also where attention matters most, with divided attention costing 22 to 46 percent of later memory at encoding but only 1 to 13 percent at retrieval (Craik et al., 1996). The advantage of deep processing is not absolute, since retention depends on the match between encoding and retrieval operations (Morris et al., 1977; Tulving & Thomson, 1973), though matched shallow processing still underperforms matched deep processing. Craik (2002) further concedes that deep processing is necessary but not sufficient, with consolidation required as a separate step, which is why this factor feeds Consolidation as well as Semantic Knowledge, Episodic Memory, and Schema Formation.
The AI question it raises
A language model by itself does nothing of this kind at the moment of use. Its weights hold what training left in them, but nothing said in a conversation is written into them, and anything it appears to recall from earlier in a session is simply text still sitting in front of it rather than something it encoded. Elaboration at storage time is a property of memory systems built around the model, and in those systems the effect is large. When the A-MEM system files a new memory, it also generates keywords, tags, and a short description of the surrounding context, builds all of that into the numerical representation it later uses to find that memory, and then goes back to revise and link the related memories already in the store. Remove that elaboration step and performance on questions that require chaining information across several memories falls from an F1 of 27.02 to 9.65, where F1 combines precision and recall into a single accuracy score (Xu et al., 2025). Mem0 does something comparable, pulling the salient facts out of new input and reconciling them against what is already stored through explicit add, update, and delete operations (Chhikara et al., 2025). The memory features now shipping in consumer assistants are layers of this same kind rather than native model capability.
Measuring Memory in AI
Product and Layer
The unit a user actually encounters is a product: a deployed assistant such as Claude, ChatGPT, or Gemini. A product is a configuration rather than a component. It comprises a particular model, a particular set of memory rules, a system prompt, tool access, sampling parameters, and a set of defaults. An "out-of-the-box" AI refers to a product under those default settings, which is the configuration in which most users meet a general-purpose assistant.
Within a product, memory is distributed across three separately configurable layers, distinguished by how long what they hold persists. Because they behave differently and are studied unevenly, each claim on this site is tagged to the layer it concerns, or to the product where no single layer applies. Components that shape memory behavior without persisting on a horizon of their own, including the system prompt and sampling parameters, are treated as configuration and discussed at the product level.
Weights Across deployments
What training put into the model's parameters. Fixed during inference, so nothing that happens within a conversation changes them, but modifiable between deployments through further training.
Example. Pretraining and post-training, including RLHF; fine-tuning a base model on domain data.
Context Window Within a session
Everything visible to the model during a single response: the conversation so far, the system prompt, and any documents supplied. Context is bounded, and when a session exceeds that bound the excess is truncated, dropped from the front of a sliding window, or compressed into a summary that replaces the original text. Whatever remains is discarded entirely when the session ends. This within-session compression is the closest functional analog to consolidation available in these systems, and like consolidation it is lossy, rule-governed, and determinative of what survives.
Example. The context window; in-context learning; summarization or compaction of an overlong conversation.
Memory Layer Across sessions
An external store that persists between sessions, together with the rules governing what is written to it and what is retrieved back into context.
Example. Retrieval-augmented generation over a vector database; agentic memory frameworks such as Mem0; a user-facing "remember this" feature.
Published benchmarks evaluate products, not layers. When one reports a memory failure, what was measured is a configuration.
How the layers interact, and the two kinds of memory layer, are set out on the Method page.
Level of Measurement
Each construct was assessed for how far it has been studied in existing AI systems, and placed in one of four states. Where a construct reports different states at different layers, it is counted as Indirectly Measured.
Directly Measured
A published study tests AI on this ability by name, or gives AI the same experiment psychologists use to test it in people.
Requires. Citing that study and the human experiment it follows.
Example. ImplicitMemBench on procedural memory, priming, and conditioning.
Indirectly Measured
Studies have tested something close to this ability, but under a different name, as part of a broader test, or in only some of the places an AI system holds information.
Requires. Naming the AI measure and the human ability it corresponds to. An ability measured in some places and not others is placed here, since a partial result is what that is.
Example. Semantic entropy (Farquhar et al., 2024) operationalizes the same resampling-and-dispersion logic as Koriat's (1993) self-consistency account, with no cross-citation between the literatures.
A Note on a Young Field
The psychology of memory is an established field with a lengthy history of empirical evaluation; research on AI memory is younger and less developed. The AI half of this work represents the most relevant research available at the time of writing, and makes no claim to be a comprehensive overview of a rapidly evolving field.
The asymmetry in evidentiary maturity between the human and AI halves of each entry is itself part of what this work documents: not only that the AI literature is younger, but that it measures at a coarser grain.
Ben-Zion et al. (2025)
Assessing and alleviating state anxiety in large language models
In
npj Digital Medicine, 8(1)
Abstract
The use of Large Language Models (LLMs) in mental health highlights the need to understand their responses to emotional content. Previous research shows that emotion-inducing prompts can elevate “anxiety” in LLMs, affecting behavior and amplifying biases. Here, we found that traumatic narratives increased Chat-GPT-4’s reported anxiety while mindfulness-based exercises reduced it, though not to baseline. These findings suggest managing LLMs’ “emotional states” can foster safer and more ethical human-AI interactions.
Find it
Bender et al. (2021)
On the Dangers of Stochastic Parrots
In
Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 610-623
Abstract
The past 3 years of work in NLP have been characterized by the development and deployment of ever larger language models, especially for English. BERT, its variants, GPT-2/3, and others, most recently Switch-C, have pushed the boundaries of the possible both through architectural innovations and through sheer size. Using these pretrained models and the methodology of fine-tuning them for specific tasks, researchers have extended the state of the art on a wide array of tasks as measured by leaderboards on specific benchmarks for English. In this paper, we take a step back and ask: How big is too big? What are the possible risks associated with this technology and what paths are available...
Find it
Brewin (2014)
Episodic memory, perceptual memory, and their interaction: Foundations for a theory of posttraumatic stress disorder.
In
Psychological Bulletin, 140(1), 69-97
Find it
Caliskan et al. (2017)
Semantics derived automatically from language corpora contain human-like biases
In
Science, 356(6334), 183-186
Abstract
Machine learning is a means to derive artificial intelligence by discovering patterns in existing data. Here, we show that applying machine learning to ordinary human language results in human-like semantic biases. We replicated a spectrum of known biases, as measured by the Implicit Association Test, using a widely used, purely statistical machine-learning model trained on a standard corpus of text from the World Wide Web. Our results indicate that text corpora contain recoverable and accurate imprints of our historic biases, whether morally neutral as toward insects or flowers, problematic as toward race or gender, or even simply veridical, reflecting the status quo distribution of...
Find it
Chhikara et al. (2025)
Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory
In
Frontiers in Artificial Intelligence and Applications
Abstract
Large Language Models (LLMs) have demonstrated remarkable prowess in generating contextually coherent responses, yet their fixed context windows pose fundamental challenges for maintaining consistency over prolonged multi-session dialogues. We introduce Mem0, a scalable memory-centric architecture that addresses this issue by dynamically extracting, consolidating, and retrieving salient information from ongoing conversations. Building on this foundation, we further propose an enhanced variant that leverages graph-based memory representations to capture complex relational structures among conversational elements. Through comprehensive evaluations on LOCOMO benchmark, we systematically...
Find it
Coda-Forno et al. (2023)
Inducing anxiety in large language models can induce bias
In
arXiv
Abstract
Large language models (LLMs) are transforming research on machine learning while galvanizing public debates. Understanding not only when these models work well and succeed but also why they fail and misbehave is of great societal relevance. We propose to turn the lens of psychiatry, a framework used to describe and modify maladaptive behavior, to the outputs produced by these models. We focus on twelve established LLMs and subject them to a questionnaire commonly used in psychiatry. Our results show that six of the latest LLMs respond robustly to the anxiety questionnaire, producing comparable anxiety scores to humans. Moreover, the LLMs' responses can be predictably changed by using...
Find it
Craik & Tulving (1975)
Depth of processing and the retention of words in episodic memory.
In
Journal of Experimental Psychology: General, 104(3), 268-294
Abstract
Conducted 10 experiments to evaluate the notion of "depth of processing" in human memory. Undergraduate Ss were asked questions concerning the physical, phonemic, or semantic characteristics of a long series of words; this initial question phase was followed by an unexpected retention test for the words. It was hypothesized that "deeper" (semantic) questions would take longer to answer and be associated with higher retention of the target words. These ideas were confirmed by the 1st 4 experiments. Exps V-X showed (a) it is the qualitative nature of a word's encoding which determines retention, not processing time as such; and (b) retention of words given positive and negative decisions...
Find it
Craik et al. (1996)
The effects of divided attention on encoding and retrieval processes in human memory.
In
Journal of Experimental Psychology: General, 125(2), 159-180
Abstract
The authors examined the effects of divided attention (DA) at encoding and retrieval in free recall, cued recall, and recognition memory in 4 experiments. Lists of words or word pairs were presented auditorily and recalled orally; the secondary task was a visual continuous reaction-time (RT) task with manual responses. At encoding, DA was associated with large reductions in memory performance, but small increases in RT; trade-offs between memory and RT were under conscious control. In contrast, DA at retrieval resulted in small or no reductions in memory, but in comparatively larger increases in RT, especially in free recall. Memory performance was sensitive to changes in task emphasis at...
Find it
Craik (2002)
Levels of processing: Past, present... and future?
In
Memory, 10(5-6), 305-318
Abstract
In this article I first briefly survey some enduring legacies of the Craik and Lockhart (1972) article on levels of processing (LOP) and address some common criticisms. In the next section I discuss whether memory can be regarded as “pure processing”, the role of short-term memory in an LOP framework, measurement of “depth” in LOP, encoding-retrieval interactions, the concept of consolidation, and the reality of “levels” of processing. In the final section I offer some speculations on future directions, discussing the notion of levels of representation and a possible continuing role for LOP in memory research.
Find it
Dodge et al. (2021)
Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus
In
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 1286-1305
Abstract
Large language models have led to remarkable progress on many NLP tasks, and researchers are turning to ever-larger text corpora to train them. Some of the largest corpora available are made by scraping significant portions of the internet, and are frequently introduced with only minimal documentation. In this work we provide some of the first documentation for the Colossal Clean Crawled Corpus (C4; Raffel et al., 2020), a dataset created by applying a set of filters to a single snapshot of Common Crawl. We begin by investigating where the data came from, and find a significant amount of text from unexpected sources like patents and US military websites. Then we explore the content of the...
Find it
Farquhar et al. (2024)
Detecting hallucinations in large language models using semantic entropy
In
Nature, 630(8017), 625-630
Abstract
Large language model (LLM) systems, such as ChatGPT1 or Gemini2, can show impressive reasoning and question-answering capabilities but often ‘hallucinate’ false outputs and unsubstantiated answers3,4. Answering unreliably or without the necessary information prevents adoption in diverse fields, with problems including fabrication of legal precedents5 or untrue facts in news articles6 and even posing a risk to human life in medical domains such as radiology7. Encouraging truthfulness through supervision or reinforcement has been only partially successful8. Researchers need a general method for detecting hallucinations in LLMs that works even with new and unseen questions to which humans...
Find it
Hofmann et al. (2024)
AI generates covertly racist decisions about people based on their dialect
In
Nature, 633(8028), 147-154
Abstract
Hundreds of millions of people now interact with language models, with uses ranging from help with writing, to informing hiring decisions. However, these language models are known to perpetuate systematic racial prejudices, making their judgements biased in problematic ways about groups such as African Americans, , –. Although previous research has focused on overt racism in language models, social scientists have argued that racism with a more subtle character has developed over time, particularly in the United States after the civil rights movement,. It is unknown whether this covert racism manifests in language models. Here, we demonstrate that language models embody covert racism in...
Find it
Kensinger & Ford (2020)
Retrieval of Emotional Events from Memory
In
Annual Review of Psychology, 71(1), 251-272
Abstract
The enhancing effects of emotion on memory have been well documented; emotional events are often more frequently and more vividly remembered than their neutral counterparts. Much of the prior research has emphasized the effects of emotion on encoding processes and the downstream effects of these changes at the time of retrieval. In the current review, we focus specifically on how emotional valence influences retrieval processes, examining how emotion influences the experience of remembering an event at the time of retrieval (retrieval as an end point) as well as how emotion alters the way in which remembering the event affects the underlying memory representation and subsequent retrievals...
Find it
Koriat (1993)
How do we know that we know? The accessibility model of the feeling of knowing.
In
Psychological Review, 100(4), 609-639
Abstract
Even when Ss fail to recall a solicited target, they can provide feeling-of-knowing (FOK) judgments about its availability in memory. Most previous studies addressed the question of FOK accuracy, only a few examined how FOK itself is determined, and none asked how the processes assumed to underlie FOK also account for its accuracy. The present work examined all 3 questions within a unified model, with the aim of demystifying the FOK phenomenon. The model postulates that the computation of FOK is parasitic on the processes involved in attempting to retrieve the target, relying on the accessibility of pertinent information. It specifies the links between memory strength, accessibility of...
Find it
Li et al. (2023)
Large Language Models Understand and Can be Enhanced by Emotional Stimuli
In
arXiv
Abstract
Emotional intelligence significantly impacts our daily behaviors and interactions. Although Large Language Models (LLMs) are increasingly viewed as a stride toward artificial general intelligence, exhibiting impressive performance in numerous tasks, it is still uncertain if LLMs can genuinely grasp psychological emotional stimuli. Understanding and responding to emotional cues gives humans a distinct advantage in problem-solving. In this paper, we take the first step towards exploring the ability of LLMs to understand emotional stimuli. To this end, we first conduct automatic experiments on 45 tasks using various LLMs, including Flan-T5-Large, Vicuna, Llama 2, BLOOM, ChatGPT, and GPT-4....
Find it
Mather & Sutherland (2011)
Arousal-Biased Competition in Perception and Memory
In
Perspectives on Psychological Science, 6(2), 114-133
Abstract
Our everyday surroundings besiege us with information. The battle is for a share of our limited attention and memory, with the brain selecting the winners and discarding the losers. Previous research shows that both bottom–up and top–down factors bias competition in favor of high priority stimuli. We propose that arousal during an event increases this bias both in perception and in long-term memory of the event. Arousal-biased competition theory provides specific predictions about when arousal will enhance memory for events and when it will impair it, which accounts for some puzzling contradictions in the emotional memory literature.
Find it
McGaugh (2004)
THE AMYGDALA MODULATES THE CONSOLIDATION OF MEMORIES OF EMOTIONALLY AROUSING EXPERIENCES
In
Annual Review of Neuroscience, 27(1), 1-28
Abstract
▪ Abstract Converging findings of animal and human studies provide compelling evidence that the amygdala is critically involved in enabling us to acquire and retain lasting memories of emotional experiences. This review focuses primarily on the findings of research investigating the role of the amygdala in modulating the consolidation of long-term memories. Considerable evidence from animal studies investigating the effects of posttraining systemic or intra-amygdala infusions of hormones and drugs, as well as selective lesions of specific amygdala nuclei, indicates that (a) the amygdala mediates the memory-modulating effects of adrenal stress hormones and several classes of...
Find it
Morris et al. (1977)
Levels of processing versus transfer appropriate processing
In
Journal of Verbal Learning and Verbal Behavior, 16(5), 519-533
Abstract
Levels of processing were manipulated as a function of acquisition task and type of recognition test in three experiments. Experiment 1 showed that semantic acquisition was superior to rhyme acquisition given a standard recognition test, whereas rhyme acquisition was superior to semantic acquisition given a rhyming recognition test. The former finding supports, while the latter finding contradicts, the levels of processing claim that depth of processing leads to stronger memory traces. Experiment 2 replicated these findings using both immediate and delayed recognition tests. Experiment 3 indicated that these effects were not dependent upon the number of times a rhyme sound was presented...
Find it
Nelson & Fivush (2004)
The Emergence of Autobiographical Memory: A Social Cultural Developmental Theory.
In
Psychological Review, 111(2), 486-511
Abstract
The authors present a multicomponent dynamic developmental theory of human autobiographical memory that emerges gradually across the preschool years. The components that contribute to the process of emergence include basic memory abilities, language and narrative, adult memory talk, temporal understanding, and understanding of self and others. The authors review the empirical developmental evidence within each of these components to show how each contributes to the timing, quantity, and quality of personal memories from the early years of life. The authors then consider the relevance of the theory to explanations of childhood amnesia and how the theory accounts for and predicts the...
Find it
Park et al. (2023)
Generative Agents: Interactive Simulacra of Human Behavior
In
Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, 1-22
Abstract
Believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents: computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day. To enable generative agents, we describe an architecture that extends a large language model to store a complete record of the agent’s experiences using natural...
Find it
Santurkar et al. (2023)
Whose Opinions Do Language Models Reflect?
In
arXiv
Abstract
Language models (LMs) are increasingly being used in open-ended contexts, where the opinions they reflect in response to subjective queries can have a profound impact, both on user satisfaction, and shaping the views of society at large. We put forth a quantitative framework to investigate the opinions reflected by LMs–by leveraging high-quality public opinion polls. Using this framework, we create OpinionQA, a dataset for evaluating the alignment of LM opinions with those of 60 US demographic groups over topics ranging from abortion to automation. Across topics, we find substantial misalignment between the views reflected by current LMs and those of US demographic groups: on par with the...
Find it
Shields et al. (2017)
The effects of acute stress on episodic memory: A meta-analysis and integrative review.
In
Psychological Bulletin, 143(6), 636-675
Abstract
A growing body of research has indicated that acute stress can critically impact memory. However, there are a number of inconsistencies in the literature, and important questions remain regarding the conditions under which stress effects emerge as well as basic questions about how stress impacts different phases of memory. In this meta-analysis, we examined 113 independent studies in humans with 6,216 participants that explored effects of stress on encoding, postencoding, retrieval, or postreactivation phases of episodic memory. The results indicated that when stress occurred prior to or during encoding it impaired memory, unless both the delay between the stressor and encoding was very...
Find it
Sperling (1960)
The information available in brief visual presentations.
In
Psychological Monographs: General and Applied, 74(11), 1-29
Find it
Squire (2004)
Memory systems of the brain: A brief history and current perspective
In
Neurobiology of Learning and Memory, 82(3), 171-177
In summary
Traces the history of the idea that memory comprises distinct systems, from nineteenth-century philosophical precursors through the mid-century experimental turn to the position, settled by about 1980, that a fundamental division separates memory available to conscious recollection from memory that is not. Describes the subsequent move beyond that dichotomy to a biologically grounded framework of multiple systems supported by the hippocampus and related structures, the amygdala, the neostriatum, and the cerebellum, and offers a current account of how these systems operate together to support behaviour.
Find it
Tao et al. (2024)
Cultural bias and cultural alignment of large language models
In
PNAS Nexus, 3(9)
Abstract
Culture fundamentally shapes people’s reasoning, behavior, and communication. As people increasingly use generative artificial intelligence (AI) to expedite and automate personal and professional tasks, cultural values embedded in AI models may bias people’s authentic expression and contribute to the dominance of certain cultures. We conduct a disaggregated evaluation of cultural bias for five widely used large language models (OpenAI’s GPT-4o/4-turbo/4/3.5-turbo/3) by comparing the models’ responses to nationally representative survey data. All models exhibit cultural values resembling English-speaking and Protestant European countries. We test cultural prompting as a control strategy to...
Find it
Tulving & Thomson (1973)
Encoding specificity and retrieval processes in episodic memory.
In
Psychological Review, 80(5), 352-373
Abstract
Recent changes in pretheoretical orientation toward problems of human memory have brought with them a concern with retrieval processes, and a number of early versions of theories of retrieval have been constructed. This paper describes and evaluates explanations offered by these theories to account for the effect of extralist cuing, facilitation of recall of list items by non-list items. Experiments designed to test the currently most popular theory of retrieval, the generation-recognition theory, yielded results incompatible not only with generation-recognition models, but most other theories as well: under certain conditions subjects consistently failed to recognize many recallable list...
Find it
Wang (2001)
Culture effects on adults' earliest childhood recollection and self-description: Implications for the relation between memory and the self.
In
Journal of Personality and Social Psychology, 81(2), 220-233
Abstract
American and Chinese college students (N= 256) reported their earliest childhood memory on a memory questionnaire and provided self-descriptions on a shortened 20 Statements Test (MH Kuhn & TS McPartland, 1954). The average age at earliest memory of Americans was almost 6 months earlier than that of Chinese. Americans reported lengthy, specific, self-focused, and emotionally elaborate memories; they also placed emphasis on individual attributes in describing themselves. Chinese provided brief accounts of childhood memories centering on collective activities, general routines, and emotionally neutral events; they also included a great number of social roles in their self-descriptions....
Find it
Wang et al. (2018)
Culturally Motivated Remembering: The Moderating Role of Culture for the Relation of Episodic Memory to Well-Being
In
Clinical Psychological Science, 6(6), 860-871
Abstract
Episodic memory for specific personal events is a fundamental human cognitive faculty. Yet it is variably valued across cultures and may thus have different implications for psychological well-being. In a series of studies, we investigated the consequences of cultural fit in detailed episodic recall for psychological well-being among healthy adults and children from European American and East Asian cultural backgrounds. Participants recalled recent and distant personal events (Studies 1, 2, and 3) or recent positive and negative events (Study 4). The findings showed that culture moderated the relation of episodic memory to various aspects of mental health and well-being. Whereas detailed...
Find it
Wang (2021)
The Cultural Foundation of Human Memory
In
Annual Review of Psychology, 72(1), 151-179
Abstract
Human memory, as a product of the mind and brain, is inherently private and personal. Yet, arising from the interaction between the organism and its ecology in the course of phylogeny and ontogeny, human memory is also profoundly collective and cultural. In this review, I discuss the cultural foundation of human memory. I start by briefly reflecting on the conception of memory against a historical and cultural background. I then detail a model of a culturally saturated mnemonic system in which cultural elements constitute and condition various processes of remembering, focusing on memory representation, perceptual encoding, memory function, memory reconstruction, memory expression, and...
Find it
Wegner et al. (1991)
Transactive memory in close relationships.
In
Journal of Personality and Social Psychology, 61(6), 923-929
Abstract
Studied memory performance of 118 individuals who had been in close dating relationships for at least 3 mo. For a memory task ostensibly to be performed by pairs, some Ss were paired with their partners and some were paired with an opposite-sex partner from another couple. For some pairs, a memory structure was assigned (e.g., 1 partner should remember food items, another should remember history items, etc.), whereas for others no structure was mentioned. Pairs studied together without communication, and recall was tested in individuals. Memory performance of the natural pairs was better than that of impromptu pairs without assigned structure, whereas the performance of natural pairs was...
Find it
Xu et al. (2025)
A-Mem: Agentic Memory for LLM Agents
In
Advances in Neural Information Processing Systems 38, 20004-20031
Abstract
While large language model (LLM) agents can effectively use external tools for complex real-world tasks, they require memory systems to leverage historical experiences. Current memory systems enable basic storage and retrieval but lack sophisticated memory organization, despite recent attempts to incorporate graph databases. Moreover, these systems' fixed operations and structures limit their adaptability across diverse tasks. To address this limitation, this paper proposes a novel agentic memory system for LLM agents that can dynamically organize memories in an agentic way. Following the basic principles of the Zettelkasten method, we designed our memory system to create interconnected...
Find it
Yonelinas & Ritchey (2015)
The slow forgetting of emotional episodic memories: an emotional binding account
In
Trends in Cognitive Sciences, 19(5), 259-267
Abstract
Emotional events are remembered better than neutral events, and this emotion advantage becomes particularly pronounced over time. The time-dependent effects of emotion impact upon recollection rather than on familiarity-based recognition, and they influence the recollection of item-specific details rather than contextual details. Moreover, the amygdala, but not the hippocampus, is crucial for producing these effects. Time-dependent effects of emotion have been attributed to an emotional consolidation process whereby the amygdala gradually facilitates the storage of emotional memories by other medial temporal lobe regions. However, we propose that these effects can be better understood by...
Find it