Line of inquiry
Inquiring lines›How do architecture and constraint…›How can systems understand user in…›this line of inquiry
Can external memory modules outperform model weight consolidation for long-term retention?
A broader line of inquiry — a family of 102 specific questions the research asks around this. Follow one into its inquiring-line page, or move sideways to a related line below.
Questions in this line of inquiry 102
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Does moving memory outside model weights avoid the limitations of in-weight retention?
- Can precomputed inferences be stored in memory modules between model interactions?
- Does including full context always degrade memory retrieval quality in practice?
- Do long-term memory modules outperform consolidation into fast weights?
- What persistent memory architectures best support storing precomputed inferences across sessions?
- How do adaptive memory modules compare to feedback-based working memory for long context?
- Do less technical users perceive memory inference differently than expert users do?
- Why do accumulated memory systems sometimes hurt continual learning?
- How often do users' memory organization needs change during different task stages?
- How should memory consolidation timing differ across multiple timescales?
- How do retention gates regularize forgetting across different sequence model architectures?
- Why do external memory consolidation systems fail worse than naive in-context learning on continual tasks?
- Why do accumulated memories hurt continual learning more than no memory?
- How does context budget create tradeoffs between memory and skills?
- Can data pruning strategies exploit the finite nature of memorization capacity?
- Do retrieval-augmented memory systems actually solve the compartmentalization problem?
- What capacity limits does the memory model face as corpus grows?
- How do different mental models of LLM memory affect user expectations?
- Can compressed long-term memory outperform fixed-window token retention?
- Can persistent memory architectures enable AI to reuse and stabilize invented concepts?
- Why does attending to own latents work better than bolted-on external memory stores?
- What makes factual memorization less efficient than tool-based retrieval?
- How do user preferences for memory controls split across different user types?
- Can native memory procedures acquired through training handle stale or incorrect cached information?
- How should memory systems split between short-term and long-term storage?
- Why does semantic memory abstraction outperform raw episodic recall for personalization?
- Why does fine-tuning for continuous space cause catastrophic forgetting?
- Can continuum memory systems prevent catastrophic forgetting in neural networks?
- What gets lost when we describe memory as retrieval?
- Why does in-weight memorization fail compared to tool-based fact access?
- Why does persistent memory alone fail to create genuine position-holding in models?
- Can understanding human memory errors improve chatbot memory system design?
- Can adaptive memory modules combine long-term filtering with short-term attention benefits?
- Why does recency-based recall outperform semantic similarity for episodic memory?
- How do fixed recurrent states trade off copying accuracy for filtering ability?
- Why do memory panels listing content miss the transparency problem users face?
- Can neural modules memorize surprising tokens as adaptive long-term memory?
- Does memory transparency require different design than memory accuracy?
- How does in-weight memorization scale with model parameter count?
- Can the same model appear to remember differently across different products?
- How do complementary learning systems explain the need for fast and slow consolidation?
- Why does semantic deduplication reduce memorization in fine-tuned models?
- Can fixed-size latent states losslessly store arbitrary input context?
- What mechanism transfers explicit memories into parametric model weights?
- Can episodic and semantic memory improve long-horizon task reasoning?
- Does composing multiple continual learning mechanisms reduce forgetting more than single approaches?
- Why does recall on demand not predict whether memory surfaces during user interaction?
- Do models with unfilled memorization capacity appear to generalize falsely?
- How can a forgetting policy preserve rare knowledge while preventing over-generalization?
- What happens when user preferences conflict with AI memory handling decisions?
- What mental models do people hold about how LLM memory technically works?
- What prevents users from manually policing every memory in their systems?
- Do users want granular control over when memories are generated and deleted?
- Can document repetition accidentally memorize sensitive information instead of learning?
- How do cortical columns implement local inference over memory cycles?
- How do compressed persistent memory states inside networks compare to attention for long context?
- Is forgetting in language models reversible or permanent knowledge loss?
- Why do language models ignore condensed memory even when it is the only memory?
- Does semantic memory improve AI personalization more than episodic memory?
- Why does distilling reasoning strategies outperform raw trajectory memory?
- What causes catastrophic forgetting during domain knowledge embedding?
- Why do abstract semantic memories outperform specific interaction histories for journey discovery?
- What interaction costs and task delays do proactive memory controls introduce?
- When live conversation contradicts stored memory, how do products decide what wins?
- Can external managers optimize context better than the model itself?
- Why does storing past judgments in memory make current evaluations worse?
- Can memory workspaces resolve contradictory evidence that stateless systems miss?
- When does combining episodic and semantic memory reduce personalization performance?
- Can conversational memory store precomputed thoughts instead of raw interaction history?
- How does compressing memory between iterations prevent overthinking?
- How do retrieved memories differ from decision-context passages for prediction?
- Should memorability systems rely on individual reports instead of group-level signals?
- How does distributional shift toward rare inputs change memorization reliance?
- How can memory shift from a passive datastore to an actively trained component?
- Does state persistence in AI systems create the same temporal presence as human waiting?
- How much does memorization capacity limit a model's ability to learn new information?
- Why do CoALA and Letta disagree on what counts as working memory?
- What makes structured memory schemas more stable than freeform text summaries?
- Can offline recurrent passes replicate sleep-based memory consolidation in AI?
- How do the six memory components combine across explicit and implicit paths?
- Which benchmarks benefit most from adding a separate memory module?
- How does stored memory actually get selected and inserted into a given request?
- Can compressive memory track what matters most across 35 conversation sessions?
- Can a memory module be swapped between different base models?
- What makes memorized paragraphs harder to corrupt than generic text?
- What triggers control processes to act on stored preference knowledge?
- Why are rare tokens the hooks for verbatim model memorization?
- What is the theoretical capacity limit before memorization saturates?
- What makes a memory reachable in the right context?
- What memory and planning capabilities do AI companions need for evolving user needs?
- How should memory systems handle deletion as a structural property?
- Why do hybrid memory systems outperform single-tier AI architectures?
- Where does institutional erosion of oversight differ from individual memory gaps?
- What counts as genuine memory under the Extended Mind thesis?
- What classifier accuracy is needed to assign memory roles reliably at retrieval time?
- What makes memory trajectories topologically stable under persistent reuse?
- How does continuous implicit memory formation differ from explicit memory encoding?
- Why does connectivity between memory modules matter more than storage capacity?
- How do the three grokking phases connect to memorization capacity limits?
- Can vector store deletion truly prevent information recovery?
- What distinguishes memory retrieval failures from failures to act on retrieved memory?
- How should governance apply to memory that emerges in ordinary infrastructure?