selective prediction
- Papers
- 3
- Released code
- 0
- First seen
- Sept 2026
- Latest
- Oct 2026
3 papers in the last two months, against 0 in the two before.
Start here
Two papers to read first, in the order they came out- 1Task-aware certificates safely automate part of LLM agent evaluation
A task-level bootstrap certificate ensures LLM judge error stays under budget, enabling safe selective automation of agent evaluation.
Sept 2026 - 2A modular abstention method separates accuracy training from abstention decisions
Training for accuracy then using a separate confidence estimator for abstention achieves strong and controllable abstention performance.
Oct 2026
How fast it is moving
Often studied with
confidence calibrationConfidence calibration aligns a model's predicted confidence with its actual accuracy across predictions.3 togetheragent safetyRisks from AI agents including timing errors, shutdown avoidance, collusion, overclaiming, and unsafe execution, mitigated by constrained decoding, safety harnesses, and swarm governance.1 togetheragentic evaluationMeasuring AI agents' performance in complex tasks, highlighting failures in planning, monitoring, and real1 togetherWho is working on it
The papers
Most central to this idea first, not most recent.
- Research Labcs.CL
A modular abstention method separates accuracy training from abstention decisions
Training for accuracy then using a separate confidence estimator for abstention achieves strong and controllable abstention performance.
University of Chinese Academy of Sciences, State Key Laboratory of AI Safety, Institute of Computing Technology, CAS · Oct 2026
- Research Labcs.CL
Task-aware certificates safely automate part of LLM agent evaluation
A task-level bootstrap certificate ensures LLM judge error stays under budget, enabling safe selective automation of agent evaluation.
Independent Researcher, University of Chinese Academy of Sciences · Sept 2026
- Top Universitycs.AI
Consequence-aware execution reduces harmful effects of tool-using AI agents
Allocating scrutiny by consequence-weighted risk and reversibility roughly halves harmful side effects of tool-using agents compared with uncertainty ranking, in controlled simulation.
Carnegie Mellon University, Feinity LLC · Oct 2026
Concepts are extracted from each paper and reused across the corpus, so this page grows on its own as the desk reads.