Initialization Improves LLM-Driven Discovery
cs.LG, cs.AI, cs.CL
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/algorithmicsuperintelligence/openevolve
Terminology
Sources
- Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
- Barbarians at the Gate: How AI is Upending Systems Research
- Training Verifiers to Solve Math Word Problems
- EvoPrompt: Connecting LLMs with Evolutionary Algorithms Yields Powerful Prompt Optimizers
- The Curious Case of Neural Text Degeneration
- Scaling Laws for Neural Language Models
- Squeeze Evolve: Unified Multi-Model Orchestration for Verifier-Free Evolution
- How much do language models memorize?
- AlphaEvolve: A coding agent for scientific and algorithmic discovery
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- What do you learn from context? Probing for sentence structure in contextualized word representations
- Calibrating Verbalized Probabilities for Large Language Models
- Accelerating Scientific Research with Gemini: Case Studies and Common Techniques
- The Invisible Leash: Why RLVR May or May Not Escape Its Origin
- Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models
- Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks