Supersede: Memory-Update Gap in LLM Agents - a vedant33 Collection

vedant33 's Collections

Supersede: Memory-Update Gap in LLM Agents

Supersede: Memory-Update Gap in LLM Agents

updated Jun 30

Open RL environment where the reward is temporal fact-currency. GRPO-trained Qwen2.5-3B LoRA lifts held-out supersession 9.0 -> 16.7 percent.

Supersede: Diagnosing and Training the Memory-Update Gap in LLM Agents

Paper • 2606.27472 • Published Jun 25

Note Paper
vedant33/supersede-qwen2.5-3b-grpo-lora

Text Generation • Updated Jun 30 • 16

Note GRPO LoRA adapter (9.0 -> 16.7%)
vedant33/supersede-rl-episodes

Viewer • Updated Jun 30 • 2.2k • 62

Note RL episodes (train + held-out test)
Sleeping

Agents

Supersede Supersession Demo

🧠

Supersession demo, GRPO-trained Qwen2.5-3B

Note Interactive supersession demo