cd /news/large-language-models/when-should-llm-agents-forget-memory… · home › topics › large-language-models › article
[ARTICLE · art-170283] src=arxiv.org ↗ pub= topic=large-language-models verified=true sentiment=· neutral

When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors

A new arXiv paper (2611.17338v1) reports that hard-filtering stale facts from LLM memory systems can cut downstream accuracy when resolver errors delete query-critical facts, with a controlled sensitivity intervention across three models moving hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. On a frozen 102/102 calibration/test split, a deployment-matched gate reached 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection. The authors conclude deletion should be treated as cost-sensitive — validating query-critical false-drop risk on matched data or defaulting to label-free retention when such validation is unavailable.

by read1 min views1 publishedNov 16, 2026

arXiv:2611.17338v1 Announce Type: new Abstract: LLM memory systems often delete facts classified as stale. We show that this hard-filtering step can reduce downstream accuracy when resolver errors delete query-critical facts. Across three models, a controlled sensitivity intervention moves hard-drop from gains of +0.186 to +0.230 at perfect resolution to losses of -0.176 to -0.181 at the lowest tested reliability. Paired targeted interventions show that false-drop causes significantly greater degradation than false-keep, while a non-deleting baseline can still fail when incorrect labels steer generation. On a frozen 102/102 calibration/test split, a simple deployment-matched gate reaches 0.655 accuracy versus 0.641 for always hard-drop and 0.662 for test-informed oracle selection. Global-accuracy crossovers require gold labels and are model- and task-specific diagnostics, not deployment thresholds. Deletion should instead be treated as cost-sensitive: validate query-critical false-drop risk on matched data, or default to label-free retention when such validation is unavailable.

── more in #large-language-models 4 stories · sorted by recency
── more on @arxiv 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/when-should-llm-agen…] indexed:0 read:1min 2026-11-16 · —