Line of inquiry
Inquiring lines›How do training data and architect…›How do systems prioritize meaningf…›this line of inquiry
Why does consolidated memory sometimes hurt continual learning more than no memory?
A broader line of inquiry — a family of 53 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 53
Specific inquiring lines the field asks around this — ordered from the most general framing down to the most specific angle.
- Why does LLM memory consolidation regress below no-memory baselines?
- What makes memory consolidation fragile compared to raw trajectory storage?
- Why does uniform memory consolidation sometimes degrade below the no-memory baseline?
- Why do external memory consolidation systems fail worse than naive in-context learning on continual tasks?
- Why do accumulated memory systems sometimes hurt continual learning?
- Why do accumulated memories hurt continual learning more than no memory?
- Why do continuously consolidated agent memories eventually degrade below no-memory baseline?
- What persistent memory architectures best support storing precomputed inferences across sessions?
- How should memory consolidation timing differ across multiple timescales?
- Does compressing all past memories into one representation lose irretrievable details?
- When does memory consolidation help agents instead of hurting performance?
- Why does memory consolidation degrade agent performance below baseline?
- How should memory systems split between short-term and long-term storage?
- Why does consolidating more state sometimes hurt performance below the no-memory baseline?
- Do retrieval-augmented memory systems actually solve the compartmentalization problem?
- What makes naive memory consolidation regress below having no memory at all?
- Why does persistent memory alone fail to create genuine position-holding in models?
- What gets lost when we describe memory as retrieval?
- What makes a learned consolidation rule lossy and where does contamination enter?
- Can episodic raw memory outperform consolidated summaries in practice?
- Why do language models ignore condensed memory even when it is the only memory?
- Why does recency-based recall outperform semantic similarity for episodic memory?
- How does consolidation schedule order affect final memory quality?
- Why should consolidation be scheduled offline rather than during forward passes?
- Why is consolidation quality the binding constraint in neural memory systems?
- Why does storing past judgments in memory make current evaluations worse?
- How does compressing memory between iterations prevent overthinking?
- Why does higher agent recall make forgetting problems harder?
- Why does distilling reasoning strategies outperform raw trajectory memory?
- Why does recall on demand not predict whether memory surfaces during user interaction?
- What separates artifact recall from persistent memory commitment in agents?
- What makes memory curation harder to solve than simply expanding storage?
- Can compressive memory track what matters most across 35 conversation sessions?
- Why do CoALA and Letta disagree on what counts as working memory?
- Can offline recurrent passes replicate sleep-based memory consolidation in AI?
- Can memory consolidation fragility be detected and reversed during execution?
- What makes structured memory schemas more stable than freeform text summaries?
- What mechanisms cause aggregated group memory to diverge from group emotional displays?
- How does raw conversation history differ from distilled memory profiles?
- How does continuous implicit memory formation differ from explicit memory encoding?
- What computational costs does closed-loop memory refinement introduce?
- How should memory systems handle deletion as a structural property?
- Where does institutional erosion of oversight differ from individual memory gaps?
- How much actionable detail does condensation strip from raw experience?
- What makes memory trajectories topologically stable under persistent reuse?
- Why does connectivity between memory modules matter more than storage capacity?
- What makes knowledge seeding equivalent to hippocampal replay in the brain?
- Why do agents ignore condensed experience even when it is the only evidence available?
- How does co-activation shape which memories become linked together?
- Can episodic and semantic memory split the same way emotion and facts do?
- What distinguishes memory retrieval failures from failures to act on retrieved memory?
- How does the hippocampus bind disparate elements without storing everything itself?
- How should governance apply to memory that emerges in ordinary infrastructure?