Keep CALM and Explore: Language Models for Action Generation in Text-based Games
Open accessAbstract
Text-based games present a unique challenge for autonomous agents to operate in natural language and handle enormous action spaces.In this paper, we propose the Contextual Action Language Model (CALM) to generate a compact set of action candidates at each game state.Our key insight is to train language models on human gameplay, where people demonstrate linguistic priors and a general game sense for promising actions conditioned on game history.We combine CALM with a reinforcement learning agent which re-ranks the generated action candidates to maximize ingame rewards.We evaluate our approach using the Jericho benchmark (Hausknecht et al., 2019a), on games unseen by CALM during training.Our method obtains a 69% relative improvement in average game score over the previous state-of-the-art model.Surprisingly, on half of these games, CALM is competitive with or better than other models that have access to ground truth admissible actions.* * Code and data are available at https://github. com/princeton-nlp/calm-textgame.Observation: You are in the living room.There is a doorway to the east, a wooden door with strange gothic lettering to the west, which appears to be nailed shut, a trophy case, and a large oriental rug in the center of the room.You are carrying: A brass lantern . . .
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Work at the time of GPT-2 needed to combine GPT-2 with reinforcement learning for action generation.
“indeed, work at that time needed to combine GPT-2 with reinforcement learning for action generation (Yao et al., 2020)”
From Cognitive Architectures for Language Agents (2023) · cites this paper as Yao et al., 2020
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 48,293 · see citing works
Bag of Tricks for Efficient Text Classification
Joulin, A., Grave, E., Bojanowski, P., & Mikolov, T. (2017) · Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers · doi:10.18653/v1/e17-2068 · Cited by 4,545 · see citing works
Flagged first, then by citations
Research works printed but not linked: 10
- Prithviraj Ammanabrolu, William Broniec, Alex Mueller, Jeremy Paul, and Mark O Riedl. 2019. To- ward automated quest generation in text-adventure games. arXiv preprint arXiv:1909.06283. · arXiv
- Prithviraj Ammanabrolu and Matthew Hausknecht. 2020. Graph constrained reinforcement learning for natural language action spaces. arXiv preprint arXiv:2001.08837. · arXiv
- Prithviraj Ammanabrolu and Mark Riedl. 2019. Play- ing text-adventure games with graph-based deep re- inforcement learning. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 3557-3565.
- Hongshen Chen, Xiaorui Liu, Dawei Yin, and Jiliang Tang. 2017. A survey on dialogue systems: Re- cent advances and new frontiers. Acm Sigkdd Ex- plorations Newsletter, 19(2):25-35.
- Kyunghyun Cho, Bart van Merriënboer, Dzmitry Bah- danau, and Yoshua Bengio. 2014. On the properties of neural machine translation: Encoder-decoder ap- proaches. In Proceedings of SSST-8, Eighth Work- shop on Syntax, Semantics and Structure in Statisti- cal Translation, pages 103-111, Doha, Qatar. Asso- ciation for Computational Linguistics.
- Marc-Alexandre Côté, Ákos Kádár, Xingdi Yuan, Ben Kybartas, Tavian Barnes, Emery Fine, James Moore, Matthew Hausknecht, Layla El Asri, Mahmoud Adada, et al. 2018. Textworld: A learning environ- ment for text-based games. In Workshop on Com- puter Games, pages 41-75. Springer.
- Nancy Fulda, Daniel Ricks, Ben Murdoch, and David Wingate. 2017. What can you do with a rock? af- fordance extraction via word embeddings. CoRR, abs/1703.03429. · arXiv
- Matthew Hausknecht, Prithviraj Ammanabrolu, Marc- Alexandre Côté, and Xingdi Yuan. 2019a. Interac- tive fiction games: A colossal adventure. CoRR, abs/1909.05398. · arXiv
- Matthew Hausknecht, Ricky Loynd, Greg Yang, Adith Swaminathan, and Jason D Williams. 2019b. Nail: A general interactive fiction agent. arXiv preprint arXiv:1902.04259. · arXiv
- Ji He, Jianshu Chen, Xiaodong He, Jianfeng Gao, Li- hong Li, Li Deng, and Mari Ostendorf. 2015. Deep reinforcement learning with a natural language ac- tion space. arXiv preprint arXiv:1511.04636. · arXiv
Cited by
Sorted by citations
Survey on large language model-enhanced reinforcement learning: Concept, taxonomy, and methods
Cao, Y., Zhao, H., Cheng, Y., Shu, T., Chen, Y., Liu, G., Liang, G., Zhao, J., Yan, J., & Li, Y. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3497992 · Cited by 174 · see citing works
Cognitive Architectures for Language Agents
Sumers, T. R., Yao, S., Narasimhan, K., & Griffiths, T. L. (2023) · arXiv · doi:10.48550/arxiv.2309.02427 · Cited by 72 · see citing works
“…et al., 2019) would not support LLM agents — indeed, work at that time needed to combine GPT-2 with reinforcement learning for action generation (Yao et al., 2020); GPT-3 (Brown et al., 2020) unlocked flexible few-shot and zero-shot reasoning for NLP tasks; while only GPT-4 (OpenAI, 2023a)…”
BackgroundSelf-citationScienceWorld: Is your Agent Smarter than a 5th Grader?
Wang, R., Jansen, P., Côté, M.-A., & Ammanabrolu, P. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.775 · Cited by 45 · see citing works
Liu, J., Hao, W., Cheng, K., & Jin, D. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.110319 · Cited by 5 · see citing works
AriGraph: Learning Knowledge Graph World Models with Episodic Memory for LLM Agents
Anokhin, P., Semenov, N., Sorokin, A., Evseev, D., Kravchenko, A., Burtsev, M., & Burnaev, E. (2024) · arXiv · doi:10.48550/arxiv.2407.04363 · Cited by 8 · see citing works
“…al. , 2018; Hausknecht et al. , 2019; Shridhar et al. , 2021; Wang et al. , 2022 ] were originally designed to evaluate reinforcement learning (RL) agents [Guo et al. , 2020; Yao et al. , 2020; Ammanabrolu et al. , 2020; Ammanabrolu and Hausknecht, 2020; Tuli et al. , 2022; Adhikari et al. , 2021].”
PLM-based World Models for Text-based Games
Kim, M., Jung, Y., Lee, D., & Hwang, S.-W. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.86 · Cited by 4 · see citing works
BAGEL: Bootstrapping Agents by Guiding Exploration with Language
Murty, S., Manning, C., Shaw, P., Joshi, M., & Lee, K. (2024) · arXiv · doi:10.48550/arxiv.2403.08140 · Cited by 2 · see citing works
- Authors
- Yao, S., Rao, R., Hausknecht, M., & Narasimhan, K.
- Affiliations
- Microsoft (United States) (US)
- Princeton University (US)
- Venue
- Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- Published
- 2020
- Pages
- 8736-8754
- OpenAlex
- W3099204253
- Cited by
- 75
- References
- 2
- Citations by year
- 2026: 4 · 2025: 11 · 2024: 14 · 2023: 18 · 2022: 15 · 2021: 13
- FWCI
- 4.39
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Conference on Empirical Methods in Natural Language Processing (EMNLP) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Princeton University
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Computer Science › Artificial Intelligence › AI-based Problem Solving and Planning
- Keywords
- text-based games, action generation, language models, reinforcement learning agents, game-playing agents, large action spaces
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
