Agenda-based Narrative Extraction: Steering Pathfinding Algorithms with Large Language Models
Authors: Brian Felipe Keith-Norambuena, Carolina Inés Rojas-Córdova, Claudio Juvenal Meneses-Villegas, Elizabeth Johanna Lam-Esquenazi, Angélica María Flores-Bustos, Ignacio Alejandro Molina-Villablanca, Joshua Emanuel Leyton-Vallejos
Organizations: Department of Computing and Systems Engineering, Universidad Católica del Norte, Antofagasta, Chile · Department of Industrial Engineering, Universidad Católica del Norte, Antofagasta, Chile · Department of Chemical and Environmental Engineering, Universidad Católica del Norte, Antofagasta, Chile
Existing narrative extraction methods face a trade-off between coherence, interactivity, and multi-storyline support. Narrative Maps supports rich interaction and generates multiple storylines as a byproduct of its coverage constraints, though this comes at the cost of individual path coherence. Narrative Trails achieves high coherence through maximum capacity path optimization but provides no mechanism for user guidance or multiple perspectives. We introduce agenda-based narrative extraction, a method that bridges this gap by integrating large language models into the Narrative Trails pathfinding process to steer storyline construction toward user-specified perspectives. Our approach uses an LLM at each step to rank candidate documents based on their alignment with a given agenda while maintaining narrative coherence. Running the algorithm with different agendas yields different storylines through the same corpus. We evaluated our approach on a news article corpus using LLM judges with Claude Opus 4.5 and GPT 5.1, measuring both coherence and agenda alignment across 64 endpoint pairs and 6 agendas. LLM-driven steering achieves 9.9% higher alignment than keyword matching on semantic agendas (p=0.017), with 13.3% improvement on Regime Crackdown specifically (p=0.037), while keyword matching remains competitive on agendas with literal keyword overlap. The coherence cost is minimal: LLM steering reduces coherence by only 2.2% compared to the agenda-agnostic baseline. Counter-agendas that contradict the source material score uniformly low (2.2-2.5) across all methods, confirming that steering cannot fabricate unsupported narratives.
Figures & tables
Algorithm 1 Agenda-Driven Narrative Extraction (Based on Narrative Trails)
Overall
Alignment by Agenda Type
Method
Coherence
Alignment
Path Length
Literal
Semantic
Counter
Maximum Capacity
6.33
4.48
10.8
—
—
—
Keyword Matching
6.28
4.64
11.8
6.91
4.78
2.24
LLM-Driven ( k=5 )
6.19
4.80
11.7
6.65
5.25
2.49
Table 1: Comparison of pathfinding methods across all agendas. Coherence and alignment scores are means across both judges (1-10 scale). Path length is the average number of articles in extracted narratives. Alignment by agenda type is shown for agenda-driven methods only.
Figure 1: Coherence vs. alignment percentage differences with respect to the maximum capacity baseline for the keyword-based and LLM-based methods.
Figure 2: Narrative trails visualization showing how different agendas steer paths through the 2D embedding space from a trial run of the algorithm. The neutral baseline (optimal bottleneck coherence) is compared against four agenda-driven paths. Endpoints and shared nodes are highlighted. Background contours show article density in the UMAP projection space, illustrating that different agendas steer paths through different topical regions of the corpus—for instance, the Regime Crackdown path traverses denser clusters of government-response articles, while Freedom Uprising follows a trajectory through protest-coverage clusters.
Figure 3: Equivalent narrative map showing the structure of extracted storylines from a trial run of the algorithm. All five paths (neutral baseline plus four agendas) flow from the source article at top to the target at bottom. Higher resolution version available here: https://figshare.com/s/1aff7f3727b7ad4fa9dc
Understanding multi-perspective alternative narratives requires identifying how their information agrees, conflicts, or differs across sources. Existing work on cross-text relations largely focuses on categorizing relations between predefined text pairs, such as entailment or contradiction, rather than directly extracting such information from full narratives. To address this gap, we introduce Overlap-Unique-Conflict (OUC) extraction, a cross-narrative task that extracts all overlapping, conflicting, and unique clauses from two narratives. To support this study, we construct a benchmark of approximately 22K narrative pairs and 140K OUC instances spanning factual, argumentative, and political discourse. Evaluating 14 open-source LLMs (0.6B-35B), we find that unique information is far easier to extract than overlap and conflict: the strongest model, Gemma-4-31B, reaches only 61.13% F1-score on overlap and 48.58% on conflict, against more than 75% on unique. Further diagnostic analysis reveals that this difficulty does not stem from relation recognition alone, but rather from a failure to pair and extract the corresponding clauses from full narratives, especially in smaller models. Nevertheless, learning these extractions with task-specific supervision narrows the gap considerably: a fine-tuned Qwen-3-8B gains 15-28% absolute over its baseline and surpasses models roughly four times its size (e.g., Qwen-3.6-35B) on several tasks. Even so, overlap and conflict remain well below satisfactory, leaving cross-narrative clause extraction an open challenge.
Although large language models (LLMs) have demonstrated impressive creative fiction generation, they struggle to maintain narrative consistency and coherent plot lines in long-form stories. In this work, we introduce a unified framework for long-form narrative generation and verification. MAGNET, a multi-agent goal-driven narrative engine for storytelling, generates stories with persona-grounded character agents that propose actions based on a shared world state and evolving story goals, while ATLAS is a graph-based pipeline that compares scene-level world representations across a generated story to detect hallucinations. By evaluating MAGNET using an LLM editor, pairwise rubric scoring, and ATLAS, we show that our framework produces coherent narratives compared to single-model prompting and IBSEN. At 100 pages, MAGNET reduced annotations and hallucinations by 41 and 50%, respectively, compared to the single model baseline and by 34 and 45%, respectively, compared to IBSEN, with pairwise rubric evaluation showing similar results. These results suggest that long-form narratives can emerge from explicit world-state tracking and goal-driven multi-agent generation, providing a foundation for controllable and structurally coherent long-form narrative generation.
Aayush Aluru, Chloe Ho, Muhammad Hammouri +5
Pocket FM · Princeton University · University of Michigan +2
Long Narrative Reasoning is an essential capability for processing and reasoning over complex narratives. While retrieval-augmented generation provides a promising framework, existing methods still face two critical challenges: cognitive islanding and cross-layer evidence disconnection. To address these issues, we propose PonsRAG, a coordinated RAG framework inspired by the biological pons. PonsRAG consists of two key components: Triple-Layer Indexing, which organizes documents into a connected knowledge structure to bridge cognitive islands, and Coordinated Reasoning, which retrieves evidence across distinct layers and integrates cross-layer information into a unified context. We evaluate PonsRAG on four long-context narrative benchmarks, and experimental results show that it outperforms the strongest baseline, achieving a 11.56% relative improvement in average accuracy on multi-choice tasks.
Rongchen Zhao, Yu Chen, Juyuan Wang +4
School of Software Engineering, Sun Yat-sen University · School of Future Technology, South China University of Technology · TikTok Inc +1