CiteArk

运行记录

6

每次复现都分别记录 Execution 是否完成，以及 Evidence 对 Claim 的 Assessment 结论；两种状态不会互相替代。

全部6执行成功5支持 Claim0挑战 Claim0反驳 Claim0证据不足6运行中0排队中0执行失败或中断0
pipeline_202609080部分执行无法判定—

Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall · claim_fig8_asymmetric_supervision_stages · L4

google/gemini-3.8-flashcommit 0ea097947enode urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact a3d72cf667

pipeline_202609070执行成功无法判定—

Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality · claim-downstream-zeroshot-mamba2-130m · L4

google/gemini-3.8-flashcommit 52b432caf9node urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact d800773a8e

pipeline_202609070执行成功无法判定—

Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality · claim-ssd-speedup-over-scan · L4

google/gemini-3.8-flashcommit fad496d8bdnode urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact a9d748b5cf

pipeline_202609070执行成功无法判定—

Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality · claim-downstream-zeroshot-mamba2-130m · L4

google/gemini-3.8-flashcommit 64491bf630node urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact 982534725a

pipeline_202609070执行成功无法判定—

Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality · claim-ssd-speedup-over-scan · L4

google/gemini-3.8-flashcommit b6846285c7node urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact f56eef90c6

pipeline_202609020执行成功无法判定—

LongPIBench: A Long-Context Benchmark for Prompt Injection · claim-heuristic-attacks-effective · L4

openai/gpt-5.6-lunacommit b3fdc116fenode urn:citeark:actor:464b299983a83142da28f82edfc618ff2a37357244d512f570d296891c119f6eartifact d9e72fbfe1