pith. sign in

Takeshi Kojima

Identifiers

  • name variant Takeshi Kojima 0.60 · backfill

Papers (9)

  1. OrderGrad: Optimizing Beyond the Mean with Order-Statistic Policy Gradient Estimation cs.LG · 2026 · author #6
  2. On Advantage Estimates for Max@K Policy Gradients cs.LG · 2026 · author #8
  3. Clustered Self-Assessment: A Simple yet Effective Method for Uncertainty Quantification in Large Language Models cs.CL · 2026 · author #2
  4. Zipping the Thought: When and How Compressed Reasoning Data Works in LLM Post-Training cs.AI · 2026 · author #3
  5. Understanding Emergent Misalignment via Feature Superposition Geometry cs.AI · 2026 · author #3
  6. Emergent Analogical Reasoning in Transformers cs.AI · 2026 · author #4
  7. RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs cs.AI · 2025 · author #4
  8. Robustifying Vision Transformer without Retraining from Scratch by Test-Time Class-Conditional Feature Alignment cs.CV · 2022 · author #1
  9. Large Language Models are Zero-Shot Reasoners cs.CL · 2022 · author #1

Mentions

  • 2205.11916 #1 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2206.13951 #1 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2606.06096 #6 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2606.06080 #8 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2606.03846 #2 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2605.28008 #3 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2509.21128 #4 · arxiv_oai · confidence 0.70 Takeshi Kojima
  • 2602.01992 #4 · arxiv_oai · confidence 0.70 Takeshi Kojima

Frequent Coauthors