Coding agents think ahead of time
arxiv.org96 points·by andre15silva··78 comments
Latent Programming Horizons in Coding Agents
arxiv.org3 points·by andre15silva··2 comments
Improving OpenAI Codex with Repo-Specific Context
codeset.ai1 points·by andre15silva··0 comments
Show HN: Haiku and Codeset > Raw Opus
codeset.ai1 points·by andre15silva··0 comments
We Made Haiku as Good as Opus. Improving Claude Code with Codeset
codeset.ai3 points·by andre15silva··0 comments
Randomness in Agentic Evals
arxiv.org1 points·by andre15silva··0 comments
Randomness in Agentic Evals
arxiv.org3 points·by andre15silva··0 comments
SWE-Bench Verified Is Now Live on Codeset
codeset.ai1 points·by andre15silva··0 comments
Codeset, a platform for training and evaluating agentic code models
codeset.ai1 points·by andre15silva··0 comments
Codeset
codeset.ai2 points·by andre15silva··0 comments
Gradient-Based Program Repair: Fixing Bugs in Continuous Program Spaces
arxiv.org16 points·by andre15silva··0 comments
Polkadot Naming Service
pns-swart.vercel.app2 points·by andre15silva··0 comments
RepairBench: Leaderboard of Frontier Models for Program Repair
repairbench.github.io1 points·by andre15silva··0 comments
RepairLLaMA: Efficient Representations and Fine-Tuned Adapter for Program Repair
arxiv.org2 points·by andre15silva··0 comments
GitBug-Java: A Reproducible Benchmark of Recent Java Bugs
github.com2 points·by andre15silva··1 comments