The authors frame Memory Consolidation (MC) as a common objective that unifies context-specified editing, skill learning, and behavioral steering, and build a five-task-family testbed measuring consolidation and retention under single, batch, and sequential updates.
Cortex folds context benefits into model weights without degrading other skills
Zhang et al. show that generating prospective queries and distilling context-conditioned outputs into low-rank or structural updates lets models retain context gains across sequential and batch edits.
Independent
Jintian Zhang · Feng He · Xinnian Liang · Ziwen Xu · Haoming Xu · Haojing Huang · +2 more
Research Digest··2 min read
The authors formalize Memory Consolidation as internalizing context-induced performance gains while preserving non-target capabilities, then propose Cortex, a method that anticipates future uses of context and writes them into parametric memory.
Why this paper
Independent
In one line
Cortex makes context benefits persist after removal by consolidating them into parametric model updates via prospective queries and low-rank changes.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§