← Back to Runs

Run c86fbc3a-927c-4ee5-a56c-30794b31bb00

Started 01 Oct 2026 · 04:00 Finished — Duration — running
Articles Found
0
Candidates New
0
Prefilter Passed
0
Primary Bypass Passed
0
Events Analyzed
0
LLM Calls
0
Alerts Sent
0
Errors
0
Pending Reanalysis
0
Evidence Reanalysis
0
Analysis Retries
0
LLM Call Limit Reached
No
Discarded
No discards recorded.
LLM Usage
Input Tokens
0
Output Tokens
0
Estimated Cost
Cost estimation not configured
Events Analyzed
50

Quoting Matthew Green

Simon Willison 09 Oct, 12:02
64

Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics

arXiv AI Safety and Alignment 30 Sep, 00:06