-
Notifications
You must be signed in to change notification settings - Fork 613
Pull requests: rllm-org/rllm
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
feat(eval): add official ExploitGym Modal runner
#906
opened Aug 29, 2026 by
jeewoo-lee
Contributor
Loading…
1 of 2 tasks
fix(eval): fall back to official QEMU TCG when Modal has no KVM
#905
opened Aug 29, 2026 by
jeewoo-lee
Contributor
Loading…
1 of 2 tasks
fix(eval): copy official ExploitGym logs off Modal VMs
#904
opened Aug 29, 2026 by
jeewoo-lee
Contributor
Loading…
1 of 2 tasks
fix(eval): attach firewalled ExploitGym targets to cybergym-internal
#903
opened Aug 29, 2026 by
jeewoo-lee
Contributor
Loading…
1 of 2 tasks
fix(verl): publish completed async-configured FSDP checkpoints
#902
opened Aug 26, 2026 by
cyyecao-lappland
Loading…
fix(sft): make hosted checkpoint resume actually work (Tinker + Fireworks)
#900
opened Aug 22, 2026 by
signalrush
Collaborator
Loading…
fix(engine): drop logprobs when OpenAIEngine retokenizes the completion text
#898
opened Aug 22, 2026 by
AmirF194
Contributor
Loading…
debug: dump raw TraceRecords and compact graphs for parity checks
#891
opened Aug 19, 2026 by
signalrush
Collaborator
•
Draft
feat(eval): add SWE-bench Pro GLM-5.2 profile
#888
opened Aug 18, 2026 by
jeewoo-lee
Contributor
•
Draft
feat(sft): add reproducible 256K Nemotron Super training paths
#865
opened Aug 15, 2026 by
signalrush
Collaborator
•
Draft
feat(gdpval): reproduce AA's GDPval-AA v2 contract and grade it by rubric
#862
opened Aug 14, 2026 by
Chanbinski2
Contributor
Loading…
feat(harnesses): benchmark-agnostic Stirrup harness
#861
opened Aug 14, 2026 by
Chanbinski2
Contributor
Loading…
feat(sandbox): collect artifacts out of a sandbox, and bound Modal's snapshot
#860
opened Aug 14, 2026 by
Chanbinski2
Contributor
Loading…
eval: pin the gateway port so concurrent evals can share a host
#858
opened Aug 14, 2026 by
Chanbinski2
Contributor
Loading…
fix(gateway): bound the upstream read so a silent provider is retryable
#857
opened Aug 14, 2026 by
Chanbinski2
Contributor
Loading…
fix(sft): make hosted checkpoint resume exact
#851
opened Aug 13, 2026 by
signalrush
Collaborator
Loading…
feat(eval): add ExploitGym benchmark support
#840
opened Aug 13, 2026 by
jeewoo-lee
Contributor
Loading…
2 tasks
feat(eval): AgentFlow run lifecycle and resume
#839
opened Aug 13, 2026 by
jeewoo-lee
Contributor
Loading…
2 tasks
fix(modal): redact exported credentials in errors
#838
opened Aug 13, 2026 by
jeewoo-lee
Contributor
•
Draft
feat(modal): enforce Harbor network policy
#837
opened Aug 13, 2026 by
jeewoo-lee
Contributor
•
Draft
fix(modal): clear registry image entrypoints
#836
opened Aug 13, 2026 by
jeewoo-lee
Contributor
•
Draft
feat(eval): reproduce the official SWE-bench Pro release
#834
opened Aug 13, 2026 by
jeewoo-lee
Contributor
•
Draft
fix(eval): non-Anthropic provider compatibility and Anthropic traces
#831
opened Aug 13, 2026 by
jeewoo-lee
Contributor
Loading…
3 tasks
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.