Jetriever ranks agent memories by utility not similarity, hitting gold in 84.8% of cases
_reachsumit · x · 2026-11-18
Jetriever replaces similarity search with full-pool utility scoring for agent memory retrieval, ranking each stored memory by Jev's decision score for usefulness to the current question.
- On InMind's 125 indirect-query tasks, it places the gold exchange first in 84.8% of cases; BM25 and MiniLM baselines indexing user turns reach 1.6% and 0.0%.
- A 20-question comparison: Jev and Qwen3.5 9B each find 90% of gold exchanges by rank 50 (Laya: 30%); at rank one Jev reaches 85% vs Qwen3.5 9B's 40%.
- On 60 MemoryAgentBench questions, Jetriever scores 55.4% vs ReFind's 44.3%.
- Cost: USD 0.00834 per InMind question, USD 0.03103 per MAB question.
Takeaway: world knowledge plus decision-oriented scoring makes affordable full-pool ranking viable.
More from coding & agent
- Ramen open-sources eval gates and guardrails for shipping MCP tools like code — Ok_Plum3595 · 2026-11-18
- RAG-FGO paper frames RAG query rewriting as black-box optimization over prompts and decoding — _reachsumit · 2026-11-18
- StreamEmbed: 3.7x faster screenshot embedding via pixel-diff pruning and KV reuse — _reachsumit · 2026-11-18
- Reddit debate: does having multiple AIs cross-check each other actually save time? — shopaholic88 · 2026-11-18
- As agents do the work, project management tools should track invariants, not tasks — _AustinCalvert_ · 2026-11-18
- Sol6.2 does PCB layout for ~$4 in 40 min, rivaling fable5.2 at 1/15 the cost — yacineMTB · 2026-11-18