rmems/agentic-coding-trajectories
Viewer • Updated • 3 • 593 • 1
Preserved for evaluation, provenance, quality analysis, and cross-generator research. No downstream model-training or weight-update
Note Independent: observable coding-agent planning, tool use, debugging, and correction.
Note Spikenaut: relay-gated state, proposal, decision, reward, and update trajectories.
Note Independent: same-context preference data for failure recovery, alignment, and reward modeling.
Note Spikenaut: event-stream and structured language-view grounding for neuromorphic training.
Note Independent: multi-agent coordination, critique, conflict resolution, and densification.