NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
105 results for “keyword:"specification"” in ICLR 2027. 23 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]From Sampling to Selection: Grounding LLM Reasoning with Predictive Models of Physical Systems▲3.0k2.33
[PDF]Agents Verify Their Assumptions, Not the Specification: Interpretation Lock-in in Autonomous Coding Agents▼5.3k1.33
[PDF]PACE: Parameter-Efficient Calibration for Robust Posterior Inference under Simulator Misspecification▼12k1.67
[PDF]Does My LLM Obey Me? Falsification-Based Attribution Tests for Prompt-Based Behavioral Conditioning▼1351.33
[PDF]When Fisher Information Mispredicts Prior Sensitivity: Model Misspecification, Not Noise▲3.0k2.50
[PDF]When Do Models Admit They Are Wrong? Failure Disclosure Is Unstable Under Reinforcement Learning▼1.2k1.67
[PDF]SG-LegalClarify: a Large-scale Benchmark for Legal Reasoning and Ambiguity Recognition in Singapore Law▼4.3k2.00
[PDF]fDAG-ACDC: Uncertainty-Adaptive Factor-Count Selection for High-Dimensional Causal Discovery▲1.5k3.50
[PDF]Delegation Under Test: Specification-Based Conformance Testing Across the LLM-Framework Boundary▼1.2k1.67
[PDF]The Strong, weak and benign Goodhart's law: An independence-free and paradigm-agnostic formalisation▲5.4k3.00
[PDF]Sign Betting for Misspecified Regression: Variance Adaptivity and Optimal Approximation▼4.3k2.75
[PDF]On Setting the Reference Point for the Hypervolume Indicator in Multi-Objective Bayesian Optimisation▼4.3k2.00
[PDF]Same Magnitude, Different Direction: Signed-Orientation Response in Time-Series Anomaly Detection▼5.8k2.33
[PDF]Rich Outcomes, Selective Constraints: Goal Overspecification in Offline Goal-Conditioned Learning▲9.2k2.67
[PDF]AuXSpec: Learning to Autoformalize Program Specifications via Executable Postconditions▲1.7k3.33
[PDF]BehaviorDistill: Distilling Black-Box Software via Evolving, Verifiable Specifications▲5.4k3.00
[PDF]Contra: Discovering and Qualifying Behavior-Changing Questions for Selective Clarification in LLM Code Generation▼5.8k2.33
Withdrawn or desk rejected (23)
[PDF]From First to Last Without Changing a Prediction: Observability Is Not Specification in Agent DiagnosisWithdrawn1.67
[PDF]Residual Fusion for Cold-Start Matrix Completion and Reward-Model-Guided Generation: A Case Study in Drug Response Prediction with Specification GamingDesk rejected
[PDF]Commit Before Inspect: Machine-Enforced Outcome Blinding for Scientific Analysis AgentsWithdrawn1.00