arXiv Computation and Language

First Make It Playable, Then Make It Good: Staged Interaction Learning for Small Dialogue-Game Agents

The paper introduces Qwen-GuidePlay-2B, a 2B‑parameter language model fine‑tuned for dialogue‑game interaction. The training pipeline consists of three stages: supervised fine‑tuning on successful game trajectories, weighted turn‑level fine‑tuning, and teacher‑guided fine‑tuning that corrects formatting and evaluates examples. The resulting model achieves a clemscore of 57.12 and a statscore of 42.68 on Playpen’s public validation, ranking second in the official challenge and demonstrating that careful curation can outperform more aggressive procedural methods.

arXiv Machine Learning

1d ago

Acquire, Repair, Preserve: A Diagnosis-Guided Post-Training Recipe for Small-Model Dialogue Game Agents

The paper presents a post‑training recipe for small dialogue‑game agents that involves three steps: acquiring broad game participation via supervised fine‑tuning, repairing specific local failures with turn‑local preference pairs, and preserving general capabilities. Applied to the LM Playschool Challenge, the method raises the public clemscore from 10.67 to 38.92 and the closed in‑domain score from 13.41 to 41.17 while keeping overall static performance nearly unchanged. The gains are mainly within the targeted game family, with limited improvement on out‑of‑domain clemscore.

By Nan Li

agents fine-tuning benchmarks

arXiv AI

Jul 29

CAST: Game Solvers as Turn-Level Teachers for LLM Agents

arXiv:2607. 25308v1 Announce Type: cross Abstract: Training large language models (LLMs) to act in long-horizon games is a promising step toward generalist decision-making, yet reinforcement learning with verifiable rewards (RLVR) relies on sparse final rewards that reveal little about which decisions determine success.

By Yu Wang, Yi-Kai Zhang, Wentao Shi, Ziang Ye, Yuchun Miao, Yueqing Sun, Qi Gu, Xunliang Cai, Lan-Zhe Guo, Han-Jia Ye, Fuli Feng

llms agents reinforcement-learning efficiency

arXiv AI

Jun 19

MENTOR: Reinforcement Learning via Flexible Teacher-Optimized Rewards for Tool-Use Distillation

arXiv:2510. 18383v3 Announce Type: replace-cross Abstract: Distilling the tool-use capabilities of large language models (LLMs) into small language models (SLMs) is essential for their practical application.

By ChangSu Choi, Hoyun Song, Dongyeon Kim, WooHyeon Jung, Minkyung Cho, Sunjin Park, NohHyeob Bae, Seona Yu, KyungTae Lim

llms reinforcement-learning fine-tuning efficiency benchmarks safety

arXiv AI

Aug 5

TACT: Taxonomy-Aligned Post-Training for Pedagogically Adaptive English Tutoring

arXiv:2608. 03952v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to provide conversational practice for English-as-a-second-language (ESL) learners.

By Dongjie Yang, Siyan Lin, Leixian Shen, Rui Sheng, Huamin Qu, Zixin Chen

llms fine-tuning benchmarks

arXiv AI

Jun 2

Evaluating Interactive Reasoning in Large Language Models: A Hierarchical Benchmark with Executable Games

arXiv:2606. 00103v1 Announce Type: new Abstract: We introduce a multi-turn interactive framework for reasoning evaluation that treats reasoning as active evidence acquisition and belief updating.

By Mingyuan Fan, Weiguang Han, Daixin Wang, Cen Chen, Zhiqiang Zhang, Jun Zhou

llms benchmarks

arXiv AI

Jun 19

SoftSkill: Behavioral Compression for Contextual Adaptation

arXiv:2606. 20333v1 Announce Type: new Abstract: Agent skills are commonly deployed as natural-language Markdown files that encode answer policies, evidence-use habits, and task procedures.

By Xijia Tao, Yihua Teng, Xinyu Fu, Ziru Liu, Kecheng Chen, Yuzhi Zhao, Suiyun Zhang, Rui Liu, Lingpeng Kong

llms agents

arXiv AI

Aug 25

Evaluating Large Language Models for automatic analysis of teacher simulations

The paper evaluates Large Language Models for automatically analyzing responses in digital teacher simulations. Experiments compare DeBERTaV3 and Llama 3 across zero‑shot, few‑shot, and fine‑tuning settings, revealing that performance varies by characteristic and that Llama 3 consistently outperforms DeBERTaV3, especially when new characteristics must be identified. The findings suggest Llama 3 is preferable for dynamic simulation environments where teacher educators introduce new evaluation criteria.

By David de-Fitero-Dominguez, Mariano Albaladejo-Gonz\'alez, Antonio Garcia-Cabot, Eva Garcia-Lopez, Antonio Moreno-Cediel, Erin Barno, Justin Reich

llms agents fine-tuning

arXiv AI

Aug 6

Hallucinations on the Board: Tool-Augmented Evaluation of LLM Chess Commentary

arXiv:2608. 04240v1 Announce Type: cross Abstract: Superhuman game engines in domains like chess have made expert-level evaluations easily accessible, yet they communicate what is true without the natural-language explanations that make such expertise educationally useful to experts and non-experts alike.

By S. Ashwin Hebbar, Peiyao Sheng, Sewoong Oh, Pramod Viswanath

llms benchmarks safety

Hugging Face Trending Papers

Aug 25

Investigating Knowledge Transfer Across Interactive Dialogue Games

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...

llms fine-tuning

arXiv AI

Aug 12

INSIDE the Student's Mind: Jointly Modeling Latent Reasoning and Action in LLM Student Simulators

arXiv:2608. 10492v1 Announce Type: new Abstract: Large Language Model (LLM)-based simulators often reproduce observable actions but fail to capture the underlying reasoning behind them.

By Rose Niousha, Minwoo Kang, Narges Norouzi

llms safety

arXiv AI

Jun 2

MMG2Skill: Can Agents Distill In-the-Wild Guides into Self-Evolving Skills?

arXiv:2606. 01993v1 Announce Type: cross Abstract: Abundant procedural knowledge on the Web holds great potential for helping agents solve long-horizon tasks.

By Xinyu Che, Junqi Xiong, Yunfei Ge, Xinping Lei, Shihao Li, Hang Yan, Han Li, Yuanxing Zhang, Zhiqi Bai, Jinhua Hao, Ming Sun, Han Li, Jiaheng Liu

llms agents multimodal benchmarks

arXiv AI

Jun 30

SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models

arXiv:2601. 03555v3 Announce Type: replace Abstract: Training reliable tool-augmented agents remains a significant challenge, largely due to the difficulty of credit assignment in multi-step reasoning.

By Yuxuan Jiang, Francis Ferraro

llms agents reinforcement-learning benchmarks

Related stories