Call for Papers

We invite submissions to the 2nd Conference on Language Modeling (COLM).

Submission link: https://openreview.net/group?id=colmweb.org/COLM/2025/Conference

Questions can be directed to: colm-pcs@googlegroups.com

Key Dates
All deadlines are 23:59 AoE (anywhere on earth)

  • Abstract deadline: March 20, 2025
  • Full paper submission deadline: March 27, 2025
  • Rebuttal period: May 27 to June 10, 2025
  • Decision notifications: July 7, 2025
  • Conference dates: October 7-10, 2025

We consider a broad range of subject areas focused on language modeling for the second iteration of COLM. We consider the term "language model" in the broadest way. A non-exhaustive list of topics of interests includes:

  1. All about alignment: fine-tuning, instruction-tuning, reinforcement learning (with human feedback), prompt tuning, and in-context alignment
  2. All about data: pre-training data, alignment data, and synthetic data --- via manual or algorithmic analysis, curation, and generation
  3. All about evaluation: benchmarks, simulation environments, scalable oversight, evaluation protocols and metrics, human and/or machine evaluation
  4. All about societal implications: bias, equity, misuse, jobs, climate change, and beyond
  5. All about safety: security, privacy, misinformation, adversarial attacks and defenses
  6. Science of LMs: scaling laws, fundamental limitations, emergent capabilities, demystification, interpretability, complexity, training dynamics, grokking, learning theory for LMs
  7. Compute efficient LMs: distillation, compression, quantization, sample efficient methods, memory efficient methods
  8. Engineering for large LMs: distributed training and inference on different hardware setups, training dynamics, optimization instability
  9. Learning algorithms for LMs: learning, unlearning, meta learning, model mixing methods, continual learning
  10. Inference algorithms for LMs: decoding algorithms, reasoning algorithms, search algorithms, planning algorithms
  11. Human mind, brain, philosophy, laws and LMs: cognitive science, neuroscience, linguistics, psycholinguistics, philosophical, or legal perspectives on LMs
  12. LMs for everyone: multi-linguality, low-resource languages, vernacular languages, multiculturalism, value pluralism
  13. LMs and the world: factuality, retrieval-augmented LMs, knowledge models, commonsense reasoning, theory of mind, social norms, pragmatics, and world models
  14. Multimodal LMs: multimodality, perception, audio, and other modalities, both for reasoning and synthesis.
  15. LMs and embodiment: perception, action, and robotics
  16. LMs and interactions: conversation, interactive learning, and multi-agents learning
  17. LMs with tools and code: integration with tools and APIs, LM-driven software engineering
  18. LMs on diverse domains and novel applications: visual LMs, code LMs, math LMs, and so forth, with extra encouragements for less studied domains or applications such as chemistry, medicine, education, database and beyond

Review Process

Submissions will be double blind: reviewers and ACs cannot see author names when conducting reviews, and authors cannot see reviewer and AC names. This means that the submission must not contain acknowledgements or any link (e.g., github) that would reveal authors' identity.

We will use OpenReview to manage submissions. The reviews and author responses will not be public initially. Submissions under review will be visible only to their assigned program committee. We will not be soliciting comments from the general public during the reviewing process. Accepted papers and their reviews will be made public after decisions are made. Discussions between reviewers and program committee members and with the authors of accepted papers will be made public. Rejected or withdrawn papers, their discussions and meta data will not be published.

Anyone who plans to submit a paper as an author or a co-author will need to create (or update) their OpenReview profile by the abstract submission deadline. The information entered in the profile is critical for ensuring that conflicts of interest are handled properly.

The program will include oral presentations and posters of accepted papers. Authors can revise their paper as many times as needed up to the paper submission deadline. Changes to the paper will not be allowed while the paper is being reviewed.

Ethics Review

Reviewers and ACs may flag submissions for ethics review. Flagged submissions will be sent to an ethics review committee for comments. Comments from ethics reviewers will be considered by the primary reviewers and AC as part of their deliberation. They will also be visible to authors, who will have an opportunity to respond. Ethics reviewers do not have the authority to reject papers, but in extreme cases papers may be rejected by the program chairs on ethical grounds, regardless of scientific quality or contribution.

Submission Instructions

Authors are asked to submit paper abstracts by the abstract submission deadline. Please make sure that all authors have an OpenReview profile with the latest information. Abstracts submitted by the abstract submission deadline must be genuine, placeholder or duplicate abstracts will be removed. Please avoid meaningfully changing your abstract and title after the abstract deadline. Such changes will likely lead to sub-optimal assignment of AC and reviewers to your paper, hurting the quality of reviewing your paper will receive. The complete paper must be submitted by the full paper submission deadline, after which the modification of submissions will be locked, except for withdrawal. Please triple-verify the list of authors – not modification of authors is possible after the submission deadline for any reason.

Paper Length and Template

There is a strict upper limit of 9 pages for the main text of the submission, with unlimited additional pages for citations.

  • Authors may use as many pages of appendices (after the bibliography) as they wish, but reviewers are not required to read the appendix.
  • Authors can add an optional ethics statement to the paper; it will not count toward the page limit, but should not be more than 1 page.
  • Authors can add an optional reproducibility statement as well, which will not count toward the page limit, but should not be more than 1 page.
  • The optional acknowledgment will not count toward the page limit, but should not be more than 1 page.

Authors are required to follow the formatting directives of the template, including the font specifications, title formatting, and margin settings.

Style Files and Templates

To prepare your submission to COLM 2024, please use the LaTeX style files provided at:

https://github.com/COLM-org/Template/archive/refs/tags/2025.zip

Guides and Policies

Conflict of Interest Policy. The COLM program committee and all submitting authors must follow the COLM conflict of interest policy (https://colmweb.org/coi-policy.html). Please consult the policy to update your OpenReview profile as requested.

Author Guidelines. Authors are expected to follow the COLM Author Guide (https://colmweb.org/AuthorGuide.html).

Code of Conduct. All COLM participants, including authors, are required to adhere to the COLM code of conduct (https://colmweb.org/CoC.html). More detailed guidance for authors, reviewers, and all other participants will be made available in due course, and participation will require acknowledging and adhering to the provided guidelines.

Code of Ethics. All participants of COLM, including the submission and reviewing process, must abide by COLM’s code of ethics (https://colmweb.org/CoE.html).

Reviewing Guidelines. Reviewers and area chairs will be asked to follow the COLM reviewing guidelines (https://colmweb.org/ReviewGuide.html).

OpenReview User Moderation Policy. New profiles created without an institutional email will go through a moderation process that can take up to two weeks. New profiles created with an institutional email will be activated automatically.