Repository Issue Activity (beta)

#599 Agent can lower val_bpb without improving the model — add an optional experiment integrity log?

#638 Prevent Margin Leaks & Surprise LLM Bills: Add a Hybrid Spend Firewall

#179 Proposal: Introduce a Project-Level Long-Term Memory File and a Guidance Agent Mechanism

#200 Is there a plan to reduce hallucinations?

#633 Now with Vision and MoE architecture

#551 1

#631 AAI

#602 Hi, how does loop forever work?

#629 Question: could a license be added to this repo?

#284 Add UCB1 Dimension-Aware Search + Experiment Memory to program.md — no code changes required

#613 Feature: Experiment Novelty Score to Reduce Repeated Search Patterns

#596 RTX 3060 fork — 27 experiments, 25% improvement (1.604 → 1.201 val_bpb)

#57 Codex doesn't seem to work?

#64 Indirect prompt injection via training output fed back to agent

#550 comm.ent

#35 Why not fixed FLOPs bugdet?

#523 Question: would you consider adding a LICENSE file?

#395 License: add separate MIT license page

#519 Request to include Ascend-compatible fork in **Notable forks**

#215 Harden downloaded dataset shard cache in prepare.py

#505 Trust verification for autonomous research agents?

#460 Copilot code review cleanup job fails deleting artifacts with 403

#501 Greyforge audit: `autoresearch` is a clean demo loop, but not a serious research architecture

#491 Applied autoresearch to Marketing Mix Modeling (MMM) — 12x improvement, beat Google Meridian

#479 kentsel dönüşünm ve istanbul