01  Call for papersFAIR Summit 2027

FAIR Summit 2027 Call for Papers

We invite submissions to the inaugural FAIR Summit, a conference focused on model training and evaluation.

We welcome paper submissions on research related to post-training, reinforcement learning, model evaluation, human and synthetic data, reward modeling, model behavior, agents, safety, and related infrastructure.

For information not listed below, please contact the FAIR Summit Program Chairs. Questions can be directed to: program-chairs.fairsummit@ai-circle.org

Please see the FAIR Summit 2027 Author Guidelines for submission instructions and policies. Authors should also review the FAIR Summit Code of Ethics and AI Use Policy before submitting.

Submit a paper →Author Guidelines →

Key dates

All deadlines are 11:59 PM Anywhere on Earth (AoE)
Abstract deadlineAoE
Paper deadlineAoE
Reviews releasedReview
Author-reviewer discussionReview
Reviewer-AC discussionReview
Final decisionsDecision
FAIR Summit 2027, October [XX to XX], 2027Summit
C/01

Subject areas

We consider a broad range of research related to model training and evaluation. A non-exhaustive list of relevant topics includes:

  • Supervised fine-tuning and instruction tuning
  • Reinforcement learning and preference optimization
  • Reward modeling and verifiable rewards
  • Human and synthetic training data
  • Data quality, curation, and data mixtures
  • Human evaluation and evaluator calibration
  • Automated evaluation and LLM-as-a-judge
  • Benchmark and evaluation design
  • Model behavior and failure analysis
  • Robustness and regression testing
  • Agent and tool-use evaluation
  • RL and agent environments
  • Safety, alignment, and adversarial evaluation
  • Multilingual and multimodal evaluation
  • Reasoning, coding, mathematics, science, and domain-specific evaluation
  • Infrastructure for post-training and evaluation
C/02

Double-blind review

Submissions will be double blind. Reviewers will not see author names while conducting reviews, and authors will not see reviewer names.

Papers posted on arXiv are permitted subject to the dual-submission policy described in the Author Guidelines.

C/03

Who should submit

We welcome submissions from researchers new to FAIR Summit and from authors working across academic, industry, and independent research settings.

Peer review requires significant time and effort from the research community. We ask authors to submit work only when they believe the central claims have been examined carefully enough to be evaluated by others.

C/04

Negative results

FAIR Summit welcomes negative results, replication studies, and work that identifies limitations or failures in existing methods. A submission does not need to report a positive result, but its claims should be supported by the evidence presented.

Please read the FAIR Summit 2027 Author Guidelines carefully before submitting.