analiza aiqbench
The paper proposes a hybrid interactive-fiction framework that mixes pre-authored story segments with LLM-generated text, using semantic similarity to keep outputs aligned with authorial intent, plus a feedback loop that iteratively rewrites prompts via in-context learning. Its claims rest on a case study and participant evaluation, so the evidence is preliminary, and it's unclear how well the approach generalizes beyond the tested setting. For AI readers, the practical takeaway is an automated prompt-refinement method for maintaining consistency in long LLM sessions, which is relevant to any application where output drift matters.
Towards Self-Evolving Narrative Systems: Real-Time Adaptation and Long-Term Optimization with LLMs
arXiv:2611.01954v1 Announce Type: new Abstract: Traditional interactive narrative systems offer strong authorial control and structural coherence but often restrict player agency. In contrast, LLM-driven narratives enable open-ended generation and greater flexibility, yet frequently suffer from inconsistency and diminished narrative control. To navigate this trade-off, we propose a hybrid narrative framework that dynamically balances pre-authored content with real-time LLM generation via semantic similarity alignment. Additionally, sustaining narrative quality over extended gameplay remains a core challenge in LLM-based storytelling. To address this, we introduce a self-evolving, feedback-driven optimization framework that leverages in-context learning to iteratively refine prompts, eliminating the need for manual prompt engineering. Through a case study and participant-based evaluation, we show that our framework enhances narrative coherence while preserving player agency, effectively aligning generated content with both authorial intent and evolving player behavior.
Abstract:Traditional interactive narrative systems offer strong authorial control and structural coherence but often restrict player agency. In contrast, LLM-driven narratives enable open-ended generation and greater flexibility, yet frequently suffer from inconsistency and diminished narrative control. To navigate this trade-off, we propose a hybrid narrative framework that dynamically balances pre-authored content with real-time LLM generation via semantic similarity alignment. Additionally, sustaining narrative quality over extended gameplay remains a core challenge in LLM-based storytelling. To address this, we introduce a self-evolving, feedback-driven optimization framework that leverages in-context learning to iteratively refine prompts, eliminating the need for manual prompt engineering. Through a case study and participant-based evaluation, we show that our framework enhances narrative coherence while preserving player agency, effectively aligning generated content with both authorial intent and evolving player behavior.
Computation and Language (cs.CL)
arXiv:2611.01954 [cs.CL]
(or arXiv:2611.01954v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2611.01954 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 223 cuvinte