Mechanisms
Memory Processes · Retrieval and Surfacing
Involuntary Retrieval
Example
Someone uses a turn of phrase your grandmother used, and you're suddenly in her kitchen. You weren't trying to remember it, but it arrived anyway.
Involuntary Retrieval in Humans
Definition
A memory arriving with no preceding retrieval attempt, triggered by a cue in the environment or in ongoing thought (Berntsen, 2010). What distinguishes it from voluntary recall is not what is remembered but how the memory is reached: voluntary recall runs a deliberate, monitored search, while involuntary retrieval is associative, proceeds with little executive control, and arrives faster than a strategically retrieved memory of the same material (Berntsen et al., 2013). Its occurrence is not random and can be controlled through what is done at encoding, arriving readily when a cue is distinctively matched to a single stored event and largely failing to arrive when the same cue has been attached to many events (Berntsen et al., 2013). Two subtypes are distinguished. Involuntary autobiographical memories return a past episode with its context and are triggered by a cue the person can identify in the large majority of cases, while involuntary semantic memories, or mind-pops, return a bare fragment of general knowledge such as a word, name, or tune, carry no accompanying context or sense of self-involvement, and usually have no cue the person can identify (Kvavilashvili & Mandler, 2004).
Neural Substrate
Involuntary and voluntary retrieval engage the same core retrieval-success network: medial temporal, ventral occipitotemporal, and ventral parietal regions, with parahippocampal gyrus, precuneus, and posterior midline activity common to both, and none of these regions differing by retrieval condition. What differs is dorsal frontal cortex. Left dorsolateral prefrontal cortex showed greater activity in the voluntary than the involuntary group, and did so equally for cues that did and did not have an associated memory and across the inter-trial interval, indicating a sustained state of strategic search rather than a trial-by-trial response (Hall et al., 2014). The stronger claim, that involuntary retrieval recruits no frontal activity at all, is consistent with the same study but rests on an underpowered post-hoc analysis within a subset of the involuntary group, which the authors present as needing confirmation.
AI at a Glance
In humans, involuntary retrieval delivers a stored memory when a cue matches it without any search being launched, arriving faster and with less executive control than deliberate recall (Berntsen et al., 2013; Hall et al., 2014). In AI systems the function is performed but not by the model: a memory layer or a product-level component places stored content into the model's input on every turn regardless of whether it was wanted, and when the user's message contains no request to remember, retrieval accuracy falls sharply for every system tested while response quality falls further still (Chang & Chen, 2026; OpenAI, 2026).
- Memory layer: the software component between stored conversation history and the model, which decides what to retain and what to place in the model's input at each turn (Chang & Chen, 2026).
- Product level: the shipped consumer application, where memory behaviour is set by product design rather than by the model or by any single retrieval component (OpenAI, 2026).
Involuntary Retrieval in AI
How Far It Has Been Measured
Measurement in AI: Indirectly Measured
AI Memory Performance: Comparison Not Possible
Where It Lives in the System
What Has Been Measured
- No study was located that tests AI on involuntary retrieval by name or runs the vigilance-task, sound-cue or diary paradigms on an AI system; the closest evidence comes from studies of unrequested recall under other names (training-data reproduction, priming, proactive preference use, memory leakage), so the level is indirect (Aerni et al., 2025).
- From training: when given ordinary requests such as writing a letter or tutorial, chat models reproduced text they had learned from the internet without being asked to, with 50-plus-character matching snippets making up about 7% of output for Gemini 1.5 up to more than 15% for Claude 3 Opus across four model families (Aerni et al., 2025).
- From training: after a new fact is learned through training updates, models show a 'priming' effect in which the new keyword intrudes into unrelated contexts after as few as 3 spaced exposures, most strongly for words the model found improbable beforehand, across PaLM-2, Gemma and Llama models (Sun et al., 2025).
- In the current conversation (the model's context window, meaning what it can see right now): on PrefEval, models had to bring up a preference stated earlier when a later request made it relevant, without being reminded; on one topic, zero-shot accuracy was 0.50 for the best model (o1-preview) at 10 turns, falling to 0.14 at 300 turns, versus 0.07 and 0.02 for Claude 3.5 Sonnet, and most of 10 models were below 10% at 10 turns (Zhao et al., 2025).
- In saved memory (stored user facts carried into later tasks): models surfaced stored personal attributes in tasks where they did not belong in up to 69% of cases, and GPT-5's rate rose from 0.1% to 9.6% as tasks increased from 1 to 40, a rough counterpart of unwanted cue-triggered recall rather than a test of it (Mireshghallah et al., 2026).
- Memory-equipped agent designs already contain a part that does this job: the generative-agents architecture stores a running record of experiences and retrieves relevant ones automatically at each step to guide behaviour, without an explicit request to remember, but that retrieval was evaluated for believability, not against human involuntary memory (Park et al., 2023).
- Comparison is not possible because there is no common scale with documented human results: the only human baseline found (median zero overlap for human writing versus up to 7.5% median for models) used online posts with no count of people, no controlled procedure, possible internet access and some plagiarism, and it measures verbatim copying rather than memories coming to mind; none of the other AI studies included people (Aerni et al., 2025).
Where a Mechanism Can Live
04 Product
The whole configuration, and the only thing a benchmark observes directly.
01
Weights
Across deployments
02
Context Window
Within a session
03
Memory Layer
Across sessions
Which of the three a result belongs to is inferred from it, not seen in it.
Read more
Dorsolateral prefrontal cortex
Cortex in the prefrontal cortex
The upper outer surface of the prefrontal cortex, on and around the middle frontal gyrus, roughly Brodmann areas 9 and 46.
The dorsolateral prefrontal cortex sits on the lateral convexity of the frontal lobe, centered on the middle frontal gyrus and covering approximately Brodmann areas 9 and 46; its borders are drawn differently by different labs and no sulcus marks them. It is densely connected with posterior parietal cortex, the dorsal striatum, and the mediodorsal thalamus. Activity here tracks working-memory delays, task switching, and rule maintenance, and it is the usual target for prefrontal transcranial stimulation. Whether it stores the contents of working memory, holds the rules for acting on them, or simply reflects task difficulty is genuinely contested, and "executive function" names the dispute rather than settling it.
Frontal lobe
Lobe in the cerebral hemisphere
The lobe in front of the central sulcus and above the lateral fissure, the largest of the four in humans.
The frontal lobe runs from the frontal pole back to the central sulcus, with the lateral fissure marking its lower boundary against the temporal lobe. Its posterior strip is the precentral gyrus, and in front of that lie premotor and supplementary motor areas; everything anterior to those is prefrontal cortex. Its lateral surface folds into superior, middle and inferior frontal gyri, its ventral surface into orbital gyri and the gyrus rectus. Frontal damage is associated with changes in planning, inhibition and social conduct, but "frontal lobe function" covers many dissociable deficits rather than one.
Medial temporal lobe
Cortex in the temporal lobe
The inner fold of the temporal lobe (hippocampus, the cortices wrapped around it, and the amygdala), removed together in the classic amnesia cases.
A regional grouping rather than a single structure: the hippocampal formation, the entorhinal, perirhinal and parahippocampal cortices that feed it, and, on some definitions, the amygdala. It became the canonical "memory region" because the surgeries and infarcts that produce dense anterograde amnesia (H.M. most famously) take out most of it at once. Whether the parts constitute one memory system or several with different jobs is an open argument, and lesions in patients often cross the boundaries that would settle it.
Parahippocampal gyrus
Cortex in the medial temporal lobe
The medial temporal gyrus below the hippocampus, running from the uncus back to the isthmus of the cingulate gyrus.
The parahippocampal gyrus forms the medial edge of the temporal lobe's ventral surface, bounded laterally by the collateral sulcus and above by the hippocampal fissure. Its anterior part carries entorhinal cortex, which supplies the hippocampus with most of its cortical input through the perforant path; its posterior part is parahippocampal cortex proper, which continues into retrosplenial cortex at the isthmus. The uncus is its hooked front tip. Posterior parahippocampal cortex contains the parahippocampal place area, activated by scenes and spatial layouts; whether that response is about scenes, space, or contextual associations more generally is debated.
Precuneus
Cortex in the parietal lobe
The medial parietal cortex in front of the parieto-occipital sulcus, above the posterior cingulate and behind the paracentral lobule.
The precuneus occupies the medial surface of the parietal lobe, bounded by the parieto-occipital sulcus behind, the marginal branch of the cingulate sulcus in front, and the subparietal sulcus below. It is densely connected to posterior cingulate and retrosplenial cortex, to the superior parietal lobule, and to the thalamus. Metabolically it is one of the most active regions of the resting brain and among the earliest to show hypometabolism in Alzheimer's disease. It is reported in episodic retrieval, mental imagery, perspective-taking and self-referential tasks; no single account of what unites those is established.
Aerni et al. (2025)
Measuring Non-Adversarial Reproduction of Training Data in Large Language Models
In
arXiv
Abstract
Large language models memorize parts of their training data. Memorizing short snippets and facts is required to answer questions about the world and to be fluent in any language. But models have also been shown to reproduce long verbatim sequences of memorized text when prompted by a motivated adversary. In this work, we investigate an intermediate regime of memorization that we call non-adversarial reproduction, where we quantify the overlap between model responses and pretraining data when responding to natural and benign prompts. For a variety of innocuous prompt categories (e.g., writing a letter or a tutorial), we show that up to 15% of the text output by popular conversational...
Find it
Berntsen (2010)
The Unbidden Past
In
Current Directions in Psychological Science, 19(3), 138-142
Abstract
Involuntary autobiographical memories are memories of personal experiences that come to mind spontaneously—that is, with no preceding attempt at retrieval. They were one of Ebbinghaus’s (1885/1964) three basic kinds of memory but have been ignored by modern cognitive psychology. Recent work suggests that involuntary memories are a basic mode of remembering that operates on the same episodic memory system as voluntary (strategic) remembering and thus follows the same rules of encoding and maintenance. Due to their associative and unplanned retrieval, involuntary memories differ from voluntary memories by being more specific, by being less relevant to one’s life story and identity, and by...
Find it
Berntsen et al. (2013)
Why am I remembering this now? Predicting the occurrence of involuntary (spontaneous) episodic memories.
In
Journal of Experimental Psychology: General, 142(2), 426-444
Abstract
Involuntary episodic memories are memories of events that come to mind spontaneously, that is, with no preceding retrieval attempts. They are common in daily life and observed in a range of clinical disorders in the form of negative, intrusive recollections or flashbacks. However, little is known about their underlying mechanisms. Here we report a series of experiments in which—for the first time—the activation of involuntary memories is controlled and predicted on the basis of manipulations done at encoding. During encoding, participants were presented with pictures of scenes paired with sounds. Both scene and sound could be either unique (derived from a category that was presented only...
Find it
Chang & Chen (2026)
When Users Don't Ask: Benchmarking Context-Driven Memory Retrieval in Conversational Agents
In
arXiv
Abstract
Large language models (LLMs) are increasingly deployed as long-horizon conversational agents, motivating growing interest in mem- ory systems. However, existing benchmarks primarily evaluate memory through QA-style probing rather than in-situ conversational usage. We introduce LOCOMO-CONV, a conversational memory benchmark derived from Lo- CoMo with four query styles: dialog, implicit, counterfactual, and composed. Across five representative memory systems, we evaluate both retrieval recall and end-to-end response quality. Our experiments show that conversational framing exposes substantial retrieval gaps over- looked by QA benchmarks, especially on implicit and composed queries, which...
Find it
Hall et al. (2014)
The Neural Basis of Involuntary Episodic Memories
In
Journal of Cognitive Neuroscience, 26(10), 2385-2399
Abstract
Voluntary episodic memories require an intentional memory search, whereas involuntary episodic memories come to mind spontaneously without conscious effort. Cognitive neuroscience has largely focused on voluntary memory, leaving the neural mechanisms of involuntary memory largely unknown. We hypothesized that, because the main difference between voluntary and involuntary memory is the controlled retrieval processes required by the former, there would be greater frontal activity for voluntary than involuntary memories. Conversely, we predicted that other components of the episodic retrieval network would be similarly engaged in the two types of memory. During encoding, all participants...
Find it
Kvavilashvili & Mandler (2004)
Out of one’s mind: A study of involuntary semantic memories
In
Cognitive Psychology, 48(1), 47-94
Abstract
The study of memories that pop into one’s mind without any conscious attempt to retrieve them began only recently. While there are some studies on involuntary autobiographical memories (e.g., Berntsen, 1996, Berntsen, 1998) research on involuntary semantic memories or mind-popping is virtually non-existent. The latter is defined as an involuntary conscious occurrence of brief items of one’s network of semantic knowledge. The recall of these items (e.g., a word, a name, a tune) is not accompanied by additional contextual information and/or involvement of self—a standard feature of involuntary autobiographical memories. The paper reports several diary and questionnaire studies which looked...
Find it
Mireshghallah, Niloofar, Neal Mangaokar, Narine Kokhlikyan, Arman Zharmagambetov, Manzil Zaheer, Saeed Mahloujifar, and Kamalika Chaudhuri (2026)
CIMemories: A Compositional Benchmark For Contextual Integrity In LLMs
In
International Conference on Learning Representations, 2026, 95513-95532
Abstract
Large Language Models (LLMs) increasingly use persistent memory from past interactions to enhance personalization and task performance. However, this memory introduces critical risks when sensitive information is revealed in inappropriate contexts. We present CIMemories, a benchmark for evaluating whether LLMs appropriately control information flow from memory based on task context. CIMemories uses synthetic user profiles with over 100 attributes per user, paired with diverse task contexts in which each attribute may be essential for some tasks but inappropriate for others. Our evaluation reveals that frontier models exhibit up to 69% attribute-level violations (leaking information...
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
Sun et al. (2025)
How new data permeates LLM knowledge and how to dilute it
In
arXiv
Abstract
Large language models learn and continually learn through the accumulation of gradient-based updates, but how individual pieces of new information affect existing knowledge, leading to both beneficial generalization and problematic hallucination, remains poorly understood. We demonstrate that when learning new information, LLMs exhibit a "priming" effect: learning a new fact can cause the model to inappropriately apply that knowledge in unrelated contexts. To systematically study this phenomenon, we introduce "Outlandish," a carefully curated dataset of 1320 diverse text samples designed to probe how new knowledge permeates through an LLM's existing knowledge base. Using this dataset, we...
Find it
Zhao et al. (2025)
Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs
In
arXiv
Abstract
Large Language Models (LLMs) are increasingly used as chatbots, yet their ability to personalize responses to user preferences remains limited. We introduce PrefEval, a benchmark for evaluating LLMs' ability to infer, memorize and adhere to user preferences in a long-context conversational setting. PrefEval comprises 3,000 manually curated user preference and query pairs spanning 20 topics. PrefEval contains user personalization or preference information in both explicit and implicit forms, and evaluates LLM performance using a generation and a classification task. With PrefEval, we evaluated the aforementioned preference following capabilities of 10 open-source and proprietary LLMs in...
Find it