Metraipolis › archive › dc_1176aa3784ee

The Temporal Tug-of-War: Visualizing and Detecting RAG Conflicts in Diffusion Models via Trajectory Variance — arXiv:2609.31684

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.31684
License (as archived)
CC0-1.0
Archived
2026-10-07T21:49:51.427Z
SHA-256 of the archived text
801492f425453a314ca02cb7e58fd2686f9d9f4f09d3d2efb4971b8e123f24c3
Stable URL of this page
https://metraipolis.com/documento/dc_1176aa3784ee

How to cite

The Temporal Tug-of-War: Visualizing and Detecting RAG Conflicts in Diffusion Models via Trajectory Variance — arXiv:2609.31684 — arXiv (en), https://arxiv.org/abs/2609.31684 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1176aa3784ee (Metraipolis archive dc_1176aa3784ee, sha256 801492f425453a314ca02cb7e58fd2686f9d9f4f09d3d2efb4971b8e123f24c3)
BibTeX
@misc{dc1176aa3784ee,
  title = {The Temporal Tug-of-War: Visualizing and Detecting RAG Conflicts in Diffusion Models via Trajectory Variance — arXiv:2609.31684},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.31684},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1176aa3784ee; sha256 801492f425453a314ca02cb7e58fd2686f9d9f4f09d3d2efb4971b8e123f24c3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31684. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-Augmented Generation (RAG) introduces a specific failure mode in discrete diffusion language models: when retrieved context contradicts parametric knowledge, the iterative denoising process becomes a visible battleground between competing knowledge sources. We identify temporal semantic divergence as an observable for detecting these conflicts and introduce the Trajectory Variance Score (TVS), a simple and interpretable measure of this divergence. TVS computes the mean pairwise cosine distance of answer embeddings across independent stochastic denoising trajectories, capturing the temporal tug of war between parametric and contextual attractors. Requiring as few as two parallel inference runs, TVS is computationally lightweight. Across four diverse datasets (Synthetic, SciQ, PopQA, and CounterFact), a simple Logistic Regression classifier using TVS achieves $70.10\%$ accuracy and $0.7647$ AUROC on LLaDA. On Dream 7B, increasing the number of trajectories from two to five improves accuracy from $63.91\%$ to $69.62\%$. More complex sequential models provide only marginal improvements over the linear classifier. Evaluation across LLaDA and Dream 7B demonstrates that conflict-induced trajectory dynamics and their key properties transfer across distinct diffusion architectures.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.31684 — 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 2026-10-07T21:27:08.363Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).