Metraipolis › archive › dc_36cd3377c8e4

Binding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial Observability — arXiv:2611.07599

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.07599
License (as archived)
CC0-1.0
Archived
2026-11-06T14:50:35.936Z
SHA-256 of the archived text
f2956d8e73d328a1dcec11d1629b18ae6662f7f237266b2f7e05ba85641ebe89
Stable URL of this page
https://metraipolis.com/documento/dc_36cd3377c8e4

How to cite

Binding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial Observability — arXiv:2611.07599 — arXiv (en), https://arxiv.org/abs/2611.07599 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_36cd3377c8e4 (Metraipolis archive dc_36cd3377c8e4, sha256 f2956d8e73d328a1dcec11d1629b18ae6662f7f237266b2f7e05ba85641ebe89)
BibTeX
@misc{dc36cd3377c8e4,
  title = {Binding Paraphrased Instructions to Encountered Objects in sLLM Agents under Partial Observability — arXiv:2611.07599},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.07599},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_36cd3377c8e4; sha256 f2956d8e73d328a1dcec11d1629b18ae6662f7f237266b2f7e05ba85641ebe89},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.07599. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reliable deployment of LLM agents requires grounding diverse user expressions in their environments. Although RL-finetuned smaller large language model (sLLM) agents rival larger prompted LLMs on embodied tasks, their robustness to instruction paraphrases remains unclear. We introduce ALFWorld-Para to evaluate this robustness under controlled paraphrases, revealing 21-52% declines in success rates (SRs), predominantly driven by object synonyms. Despite retaining knowledge of synonym referents, agents often fail to take the corresponding objects even after observing them, exposing a gap between referential knowledge and object binding. Partial observability further complicates binding, as referents may remain unseen or have already been encountered under different labels. To address this challenge, we propose BINDIntent, a training-free framework that probabilistically binds referring expressions to observed labels using semantic proximity and estimated persistence against unseen candidates. Without additional training or LLM calls, BINDIntent improves synonym-paraphrase success of RL-finetuned sLLMs by 17.5-49.1 percentage points (pp) with only 0.4 ms of CPU overhead per step. It also raises GiGPO-7B's overall SR from 61.8% to 77.4%, surpassing a prompted 24B LLM. Code and benchmark are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.07599 — 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 2026-11-06T14:21:42.628Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).