fix: recover per-token ids and logprobs from content[] to stop RL KLD corruption - #457
Closed
Dranoxgithub wants to merge 4 commits into
Closed
fix: recover per-token ids and logprobs from content[] to stop RL KLD corruption#457Dranoxgithub wants to merge 4 commits into
Dranoxgithub wants to merge 4 commits into
Cursor / Cursor Bugbot
succeeded
Jul 6, 2026 in 1m 57s
Bugbot Review
Bugbot Analysis Progress (2m 0s elapsed)
✅ Gathered PR context (3s)
✅ Completed bug detection (1m 57s)
✅ Posted analysis results (1s)
Final Result: Bugbot completed review - no issues found! ✅
Request ID: serverGenReqId_b6a776ed-daa9-46d9-b938-e36948597b82
Details
Loading