自动项目说明
notes
这里持续整理与发布「notes」相关资料。当前共有 173 个可公开访问的文件，分布在 32 个子目录中；本说明会随管理员的公开上传自动同步。
- 公开文件
- 173
- 子目录
- 32
- 最近同步
- 2026-07-15 15:57:43
重要页面
暂无可链接的公开 HTML 页面。
最新公开内容
- KernelBench: Can LLMs Write Efficient GPU Kernels?
- CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark
- ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery
- MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering
- RE-Bench: Evaluating Frontier AI R&D Capabilities of Language Model Agents Against Human Experts
- PaperBench: Evaluating AI's Ability to Replicate AI Research
figures
harness_evolve/notes/figures
文件夹
_综述-自我改进与harness演化.md
harness_evolve/notes/_综述-自我改进与harness演化.md
ID b8e8641b
ref04_absolute-zero.md
harness_evolve/notes/ref04_absolute-zero.md
ID 8bbe8bff
ref05_self-rewarding-language-models.md
harness_evolve/notes/ref05_self-rewarding-language-models.md
ID 825f8129
ref06_spin-self-play-finetuning.md
harness_evolve/notes/ref06_spin-self-play-finetuning.md
ID dab4dfce
ref07_ace-agentic-context-engineering.md
harness_evolve/notes/ref07_ace-agentic-context-engineering.md
ID 2fc33e4c
ref08_mce-meta-context-engineering.md
harness_evolve/notes/ref08_mce-meta-context-engineering.md
ID 69e6c6ff
ref09_meta-harness.md
harness_evolve/notes/ref09_meta-harness.md
ID e112cd53
ref11_scientistone.md
harness_evolve/notes/ref11_scientistone.md
ID b1ec0a1f
ref12_autodata.md
harness_evolve/notes/ref12_autodata.md
ID f994a23d
ref13_adas-automated-design-agentic-systems.md
harness_evolve/notes/ref13_adas-automated-design-agentic-systems.md
ID d0ada653
ref14_self-refine.md
harness_evolve/notes/ref14_self-refine.md
ID 26ed9621
ref15_aflow.md
harness_evolve/notes/ref15_aflow.md
ID 692b4146
ref16_stop-self-taught-optimizer.md
harness_evolve/notes/ref16_stop-self-taught-optimizer.md
ID 17146d1b
ref17_self-harness.md
harness_evolve/notes/ref17_self-harness.md
ID cb8244cd
ref18_promptbreeder.md
harness_evolve/notes/ref18_promptbreeder.md
ID c6f78fa2
ref19_gepa.md
harness_evolve/notes/ref19_gepa.md
ID 57ef5301
ref20_alphaevolve.md
harness_evolve/notes/ref20_alphaevolve.md
ID 5bd57dea
ref21_shinkaevolve.md
harness_evolve/notes/ref21_shinkaevolve.md
ID 5aad6e9d
ref22_thetaevolve.md
harness_evolve/notes/ref22_thetaevolve.md
ID 9467d3b4
ref23_darwin-godel-machine.md
harness_evolve/notes/ref23_darwin-godel-machine.md
ID 086d23ec
ref24_hyperagents.md
harness_evolve/notes/ref24_hyperagents.md
ID 28466464
ref25_learning-to-discover-at-test-time.md
harness_evolve/notes/ref25_learning-to-discover-at-test-time.md
ID c8e9c81c
ref26_epistemic-uncertainty-test-time-discovery.md
harness_evolve/notes/ref26_epistemic-uncertainty-test-time-discovery.md
ID 7cde4200
ref27_sia-self-improving-ai.md
harness_evolve/notes/ref27_sia-self-improving-ai.md
ID 2309c1c6
ref28_why-llms-arent-scientists-yet.md
harness_evolve/notes/ref28_why-llms-arent-scientists-yet.md
ID 8e32e246
ref29_early-science-acceleration-gpt5.md
harness_evolve/notes/ref29_early-science-acceleration-gpt5.md
ID 746415d1
ref30_paperbench.md
harness_evolve/notes/ref30_paperbench.md
ID fba848ea
ref31_re-bench.md
harness_evolve/notes/ref31_re-bench.md
ID fbfe523d
ref32_mle-bench.md
harness_evolve/notes/ref32_mle-bench.md
ID f876e79f
ref33_scienceagentbench.md
harness_evolve/notes/ref33_scienceagentbench.md
ID 84f06e14
ref34_core-bench.md
harness_evolve/notes/ref34_core-bench.md
ID 034d42e0
ref35_kernelbench.md
harness_evolve/notes/ref35_kernelbench.md
ID 10774e09