FrontierSmith: Synthesizing Open-Ended Coding Problems at Scale | alphaXiv
Researchers from UC Berkeley, Stanford, Princeton, MIT, and Bespoke Labs developed FrontierSmith, an automated system that synthesizes open-ended coding problems by mutating existing closed-ended...
Abstract Many real-world coding challenges are open-ended and admit no known optimal solution. Yet, recent progress in LLM coding has focused on well-defined tasks such as feature implementation, bug fixing, and competitive programming. Open-ended coding remains a weak spot for LLMs, largely because open-ended training problems are scarce and expensive to construct. Our goal is to synthesize open-ended coding problems at scale to train stronger LLM coders. We introduce FrontierSmith, an automated system for iteratively evolving open-ended problems from existing closed-ended coding tasks. Start
saved by
related reading
- ALE-Bench: A Benchmark for Long-Horizon Objective-Driven Algorithm Engineering | alphaXivalphaxiv.org
- LLM-as-a-Verifier: A General-Purpose Verification Framework | alphaXivalphaxiv.org
- Introducing FrontierCode | Cognitioncognition.ai
- Introducing FrontierCode | Cognitioncognition.com
- Automated Weak-to-Strong Researcheralignment.anthropic.com
- frontier model training methodologies | Alex Wa's Blogdjdumpling.github.io
- [2203.07814] Competition-Level Code Generation with AlphaCodearxiv.org
- FrontierSWEfrontierswe.com
- AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — Google DeepMinddeepmind.google
- Program Synthesis with Large Language Modelsarxiv.org
- SWE-1.7: Frontier Intelligence at a Fraction of the Costcognition.com
- Our Problems · Proximalproximal.ai