Driving · 2026-11-03

Creating Hybrid Interactive Narrative Systems with LLM: From Formal Description to Evaluation

Jules Clerc (Heudiasyc), Domitile Lourdeaux (Heudiasyc), Mohamed Sallak (Heudiasyc), Marc Ravaine

arXiv:2611.02807PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

The authors propose an LLM-based narrative platform to compare hybrid experience managers (EMs). Managers are interchangeable modules over a shared symbolic world model (facts, operations, goals), so each session can be represented as a common state-transition plan and evaluated via indicators of player agency, EM controllability, author intent, and robustness. They instantiate three managers (open generation, MCTS repair, multi-agent NPCs) and report trade-offs on an interactive Little Red Riding Hood scenario.

Why it matters

The paper targets the “No Yardstick” problem in interactive narrative evaluation: heterogenous LLM EMs are hard to compare because their behaviors depend on scenario-specific details and internal, non-verifiable decision-making. By forcing a shared symbolic representation with explicit preconditions/effects and plan-based functional scoring, the authors provide a common ground for evaluating and selecting among different EM architectures based on measurable properties.

Method

  • Use a shared symbolic world model (facts, operations with preconditions/effects, goals) so sessions become paths in a common state-transition representation (plans).
  • Instantiate and compare three LLM-based hybrid EM modules: open generation, MCTS-based repair when blocked, and multi-agent NPC-driven progression.
  • Evaluate each manager with four plan-level properties (player agency, EM controllability, author intent, robustness) computed as summary statistics over nsim sessions on a fixed Little Red Riding Hood scenario.

Limitation

The authors state results still need confirmation with human players, and that the comparison uses a single scenario, so it does not establish generalisability; they also note evaluation cost grows linearly with the number and length of simulated sessions because properties are computed on simulated plans rather than the whole state space.

Abstract (from arXiv)

Interactive narratives let users shape a story while an Experience Manager (EM) keeps it coherent, balancing player freedom against authorial control. Large Language Models now enable a new generation of hybrid EMs, which combine run-time generation from free-text input with symbolic narrative control. However, their heterogeneity makes them hard to compare: A manager's behaviour is generally tied to the unique and specific scenario in which it operates, and much of its decision-making takes place outside of any explicit and verifiable structure. We address this with an LLM-based narrative platform whose managers are interchangeable modules operating over a shared symbolic world model of facts, operations, and goals. Because every applied operation carries explicit preconditions and effects, each session can be analysed as a plan through a common state-transition representation and evaluated, independently of how the manager works internally, with shared indicators of four properties: player agency, EM controllability, author intent, and robustness. We implement three contrasting managers, based respectively on open generation, Monte Carlo repair, and autonomous characters, and compare them through functional analysis on an interactive Little Red Riding Hood scenario. The results show that no manager dominates on every property, giving authors and EM designers a concrete basis to select a manager according to their priorities.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First announced by arXiv: 2026-11-03 · Indexed by Pub-AI: 2026-11-03 UTC

preprint

Creating Hybrid Interactive Narrative Systems with LLM: From Formal Description to Evaluation

Jules Clerc (Heudiasyc), Domitile Lourdeaux (Heudiasyc), Mohamed Sallak (Heudiasyc), Marc Ravaine

First announced by arXiv: 2026-11-03 · Original record ↗

arxiv
2611.02807
Explicit source versions (1)
  • v1 · Date not supplied
  1. Preprint available

    Source date: 2026-11-03 · First detected: 2026-11-03 UTC · Historical baseline

    Source evidence (1)
    • Official evidence ↗ · RSS guid=oai:arXiv.org:2611.02807v1; arxiv:announce_type=new; item pubDate (official metadata)
Abstract · arxiv

Interactive narratives let users shape a story while an Experience Manager (EM) keeps it coherent, balancing player freedom against authorial control. Large Language Models now enable a new generation of hybrid EMs, which combine run-time generation from free-text input with symbolic narrative control. However, their heterogeneity makes them hard to compare: A manager's behaviour is generally tied to the unique and specific scenario in which it operates, and much of its decision-making takes place outside of any explicit and verifiable structure. We address this with an LLM-based narrative platform whose managers are interchangeable modules operating over a shared symbolic world model of facts, operations, and goals. Because every applied operation carries explicit preconditions and effects, each session can be analysed as a plan through a common state-transition representation and evaluated, independently of how the manager works internally, with shared indicators of four properties: player agency, EM controllability, author intent, and robustness. We implement three contrasting managers, based respectively on open generation, Monte Carlo repair, and autonomous characters, and compare them through functional analysis on an interactive Little Red Riding Hood scenario. The results show that no manager dominates on every property, giving authors and EM designers a concrete basis to select a manager according to their priorities.