Metraipolis › archive › dc_7ea9e7ccf750

PSD: Pseudo Self-Distillation of Memory Representation Capabilities for LLM Agents — arXiv:2609.23449

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.23449
License (as archived)
CC0-1.0
Archived
2026-10-09T06:50:35.698Z
SHA-256 of the archived text
6f7647fac6d640eb93ca74eca8277004415f039ba4b3409d6509516515c67464
Stable URL of this page
https://metraipolis.com/documento/dc_7ea9e7ccf750

How to cite

PSD: Pseudo Self-Distillation of Memory Representation Capabilities for LLM Agents — arXiv:2609.23449 — arXiv (en), https://arxiv.org/abs/2609.23449 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7ea9e7ccf750 (Metraipolis archive dc_7ea9e7ccf750, sha256 6f7647fac6d640eb93ca74eca8277004415f039ba4b3409d6509516515c67464)
BibTeX
@misc{dc7ea9e7ccf750,
  title = {PSD: Pseudo Self-Distillation of Memory Representation Capabilities for LLM Agents — arXiv:2609.23449},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.23449},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7ea9e7ccf750; sha256 6f7647fac6d640eb93ca74eca8277004415f039ba4b3409d6509516515c67464},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.23449. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Memory systems are becoming a core component of LLM agents, but constructing and maintaining memory remains expensive because it relies on repeated calls to large proprietary language models. This cost creates a major barrier to deploying memory-enhanced agents at scale. In this paper, we present Pseudo Self-Distillation (PSD), a framework that enables small language models (SLMs) to construct hierarchical memory representations by distilling behavior from a strong black-box oracle through a multi-stage training pipeline. Standard distillation methods require access to teacher logits or hidden states, which closed models do not expose. Unlike conventional self-distillation settings, where supervision is derived from a model's own predictions, sampled rollouts, or aggregated outputs, PSD enables a single-model distillation setup while channeling external oracle knowledge through the prompt. PSD uses a single small model in two roles: a teacher that sees a privileged prompt containing the oracle's answer as reference context, and a student that sees only the task prompt. The student learns to reproduce the teacher's output distribution, absorbing oracle-guided behavior into its own weights without accessing the oracle's internals. On LoCoMo, PSD-trained Qwen3-0.6B, 1.7B, and 4B match or exceed GPT-4.1-mini on downstream retrieval at a fraction of the deployment cost, with off-policy PSD achieving the strongest results across most conditions. We further show that this memory-construction capability transfers out-of-distribution to LongMemEval, despite the students being trained exclusively on LoCoMo with no exposure to LongMemEval data.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL).
Categorie arXiv: Information Retrieval (cs.IR) ; Computation and Language (cs.CL); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.23449 — 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 2026-10-09T06:48:45.124Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).