# Experimenting with imperfect LLM memory

> Source: <https://discuss.huggingface.co/t/experimenting-with-imperfect-llm-memory/180679#post_2>
> Published: 2026-09-22 04:51:38+00:00

Long story short of what the repo does:

Instead of treating memory as immutable information storage, SelMem maintains mutable lived traces:

input → semantic segmentation → selective encoding → trace → recall/reconstruction → reconsolidation → forgetting/merging

Each input is stored in two forms:

sealed archive — immutable original input, used only for auditing/evaluation;

lived memory — compressed and mutable representation used by the agent.

Lived traces contain semantic content plus attributes such as salience, affect, self-relevance, fidelity, access history and permanence. Memories can decay, become latent, merge, or be reconstructed differently when recalled.

Recall uses retrieval to select memories, then a grounding/judging layer compares the reconstruction against the trace’s semantic core. It distinguishes cases such as preservation, compression, elaboration, reframing, contradiction and unrelated recall. Repeated grounding failures can trigger corrective reconsolidation