Tag
This paper identifies two failure modes in multi-objective prompt optimization for LLM judges using textual gradients: gradient dilution during optimization and instruction interference during inference, showing that joint gradient processing loses criterion-specific information.
MOCHA introduces a multi-objective optimization method for LLM agent skills, using Chebyshev scalarization and exponential annealing to handle hard platform constraints and discover Pareto-optimal variants, achieving significant improvements over existing optimizers.
This paper proposes GESD, a procedural-oriented fairness metric that measures disparities in explanation stability across subgroups, and integrates it into a multi-objective optimization framework for jointly optimizing utility, outcome fairness, and explanation fairness.
Lean Refactor presents a retrieval-augmented agentic framework for multi-objective, controllable, and version-robust refactoring of Lean proofs, achieving significant compression and compilation-time reduction.
ToolMol is an evolutionary agentic framework that combines a multi-objective genetic algorithm with an LLM-based operator to design small-molecule drugs, achieving state-of-the-art binding affinity and drug-likeness on multiple protein targets.