Repository navigation
TimeRSI combines code-agent research and chronological forecast composition around frozen TimesFM 3.0. The repository provides the method implementation and evaluation instructions.
Execution environment: The public runner uses Docker to isolate code-agent sessions and generated candidates. Implementation.
The assets contain the complete recorded primary agent-session transcripts for all 20 official mini tasks: 821 sessions and 17,992 tool calls, including prompts, candidate code and feedback. Our review of the recorded calls found no direct retrieval of current or future evaluation targets or scores.
See the artifact index for contents and checksums, and the data-access protocol for evidence scope.
Updated at 2026-10-14 02:02 PDT.