It is suggested that imitating full trajectories helps with playability, while turn-level and teacher-guided training usually improve decision-making and increase the overall score, and small models are performant simply by using careful curation strategies rather than aggressive changes.
Abstract
We present Qwen-GuidePlay-2B, a 2B-parameter language model for dialogue-game interaction. We fine-tune Qwen3.5-2B using three steps: a) SFT on only successful game trajectories from Playpen, b) weighted turn-level SFT, and c) teacher-guided SFT. The teacher model (which is a larger model) is only used to fix formatting and evaluate examples, but does not create new gold actions. Our final model scores 57.12 clemscore and 42.68 statscore on the public Playpen validation. In the officially released challenge results, our model obtains the second-highest Playpen clemscore delta among submitted systems (which is approximately +36 over its base model). Our findings suggest that imitating full trajectories helps with playability, while turn-level and teacher-guided training usually improve decision-making and increase the overall score. Alternative procedurally heavy approaches like replay-repair and hard-example mining did not help, which suggests that small models are performant simply by using careful curation strategies rather than aggressive changes. We make available both the model and the code for reproducibility.
The results suggest that broad SFT brings most of the model's capability improvement; turn-local supervision can be effective when failure detection is precise, with observed transfer concentrated primarily within-family.
Non-player character (NPC) dialogue plays a crucial role in games. Narrative-driven video games in particular depend on NPCs to help shape the player experience by contributing to narrative, immersion, and player agency. Current technologies allow large language models (LLMs) to create dynamic, context-sensitive dialogue for NPCs, yet their impact on player experience remains underexplored. Quest of Aivengarde is a custom-built role-playing game developed as a research testbed for comparing four dialogue system designs: a static control version and three LLM-driven variants that rephrase, hybridize, or fully generate NPC dialogue. Building on previous pilot and demo studies, this paper integrates the system design and full empirical evaluation to examine how levels of generative agency affect interaction quality, immersion, and player engagement. Using mixed-methods analysis of behavioral logs and post-game surveys from 64 participants, results suggest that fully open-ended LLM dialogue fosters longer, more natural conversations and is particularly engaging for casual players. We discuss design trade-offs between narrative control and conversational freedom and propose a practical framework for selecting and combining LLM-driven dialogue approaches in future game development and research.
Emil Rimer, A. Petersen, Rasmus Ploug et al.· International Conference on...· 0 citations
Dialogue games represent a challenging setting where complex cognitive skills are required to accomplish tasks while coordinating with other players. Considering that language represents an interface for both understanding the game rules and executing actions, it is reasonable to assume that training on a specific language game will enhance specific capabilities that might be relevant for other tasks as well. Motivated by this rationale, in this paper, we investigate how knowledge transfers across different dialogue games. We study transferability by finetuning LLM models on games from the clembench suite (Chalamalasetti et al., 2023) and performing two analyses: i) we derive a task-transferability graph using a binary integer optimization program from Zamir et al. (2018), using task performance as the main metric; and ii) we compute task vectors (Ilharco et al., 2022) for each game to study similarities across finetuned models and their task transferability. In our first analysis, we find that some games benefit more from transfer than finetuning, and that the visuospatial family (e.g., exploration games) transfers best. With our task vector analysis instead, we find that similarity-based approaches capture game-role relationships but almost no transferability patterns, suggesting that more complex metrics are required.
Filippo Momentè, Mir Nafis Sharear Shopnil, Andrea Gregor de Varda et al.· 0 citations
AnthroDial is presented, a closed-loop framework that formulates anthropomorphic dialogue as a joint problem of system architecture, executable evaluation, and diagnostic alignment and shows that anthropomorphic dialogue benefits when generation, evaluation, and reward shaping share the same behavioral dimensions.
Wentao Liu, Si-Yu Song, Xi Chen et al.· arXiv.org· 0 citations
Large language models can interpret user requests plausibly yet fail to produce actions that satisfy structural constraints. We study this problem in Dungeons & Dragons (D&D) combat, where player intent must be grounded in the current game state and translated into well-formed commands for Avrae, a Discord-based D&D automation system. We fine-tune a 4-bit quantized LLaMA-3 8B model with Low-Rank Adaptation on 256 curated FIREBALL-derived examples and compare it with zero-shot, one-shot, and few-shot prompting. On 25 held-out interactions, the model achieves 56% semantic grounding accuracy and 48% strict command validity. Results suggest that parameter-efficient adaptation improves grounding, while formatting, argument, and reference errors remain common.
Ez Zwall, Kedong Chen, Yalun Zhou et al.· Proceedings of the 26th ACM...· 0 citations
User-facing tool agents must coordinate dialogue and tool use as user goals unfold over multiple turns. Yet interactive reinforcement learning typically reduces each rollout to a terminal reward, assigning the same credit to effective elicitation, errors, and later repair. The next user turn is more than context: it also provides noisy, temporally local evidence about the preceding user-to-user segment. We introduce \textbf{F}eedback-\textbf{A}ware \textbf{C}redit \textbf{A}ssignment (\textsc{FACA}), which aligns each reaction with that segment, derives a locally normalized reaction advantage, and adds it to verified terminal outcome advantage without an extra critic or rollout. Against an outcome-only Interactive GRPO control matched in simulator, visible dialogue, initialization, rollout, and optimization, \textsc{FACA} improves the nine-domain $\tau$-family average across three independently trained runs by 5.91 and 10.22 percentage points at 8B and 14B, respectively. Gains concentrate in Telecom; at 8B, randomizing reaction polarity removes the Telecom gain. The same ordering holds zero-shot on Pare-Bench and Co-Gym. These results demonstrate that next-turn user reactions provide actionable local credit for improving multi-turn user-interacting agents.
Yiwen Zhao, Zhihao Wen, Yuchen Mao et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 31, 2026
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.
