Metraipolis › archive › dc_edadf58f4aff

When Does Experience Memory Help Closed-Loop LLM Driving Agents? A Controlled Evaluation with Resampled Warmup Pools, Token-Matched Fillers, and Audited Baselines — arXiv:2610.27714

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.27714
License (as archived)
CC0-1.0
Archived
2026-10-21T07:06:47.866Z
SHA-256 of the archived text
b0d8cfb7d62ace8b9812b55ca1b939a1485b77b37e38a963294a41b4f1b7fa76
Stable URL of this page
https://metraipolis.com/documento/dc_edadf58f4aff

How to cite

When Does Experience Memory Help Closed-Loop LLM Driving Agents? A Controlled Evaluation with Resampled Warmup Pools, Token-Matched Fillers, and Audited Baselines — arXiv:2610.27714 — arXiv (en), https://arxiv.org/abs/2610.27714 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_edadf58f4aff (Metraipolis archive dc_edadf58f4aff, sha256 b0d8cfb7d62ace8b9812b55ca1b939a1485b77b37e38a963294a41b4f1b7fa76)
BibTeX
@misc{dcedadf58f4aff,
  title = {When Does Experience Memory Help Closed-Loop LLM Driving Agents? A Controlled Evaluation with Resampled Warmup Pools, Token-Matched Fillers, and Audited Baselines — arXiv:2610.27714},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.27714},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_edadf58f4aff; sha256 b0d8cfb7d62ace8b9812b55ca1b939a1485b77b37e38a963294a41b4f1b7fa76},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27714. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

When does experience memory help a closed-loop LLM driving agent? Published gains cannot say, because the two factors we find to matter most, which warmup experiences the pool contains and what any injected tokens do to the prompt, are controlled in none of them. We answer with a controlled evaluation on one audited HighwayEnv testbed (1Hz, gpt-4o-mini for the decisive contrasts). We treat the warmup pool as a resampled random factor (ten disjoint pools) and add token-matched irrelevant-filler controls, oracle-certified diagnostics, caution-pricing metrics, and quantified audits of the baseline ports. The probe suite spans flat-text retrieval (a DiLu port), a single-frame grid port (RESPOND-inspired), our episodic temporal-graph memory, a retrieval-free consolidated rulebook, and a flat K-step observation-history baseline. Four findings follow, each a property of this testbed. (1) Headline memory gains belong to the warmup draw. The reported pool is the best of ten for both retrieval memories, which beat No-Memory in only one pool of ten (sign test p=0.011), and even the rulebook's apparent pool-immunity collapsed on five fresh draws. (2) No retrieval design tested exceeds token-matched filler. The collision-only merge ceiling is reached by a single 15-token caution instruction, and metrics that price caution erase the rulebook's lead. (3) Representational sophistication is nowhere robustly operative. (4) Baseline port quality silently inverts conclusions. On this testbed, a single-pool comparison without filler controls would have reported memory gains that these controls reverse. We release the testbed, audited ports, seeds, and caches.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.27714 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-21T07:05:43.439Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).