License: ercot-website-user-agreement-eia-public-domain-apache-2.0
Oct 25 Today 17:30 v3 rules 3.2: corrected settlement and hindsight bound, refilled packs, and the rules-3.2 re-pilot (part 2) 17:30 v3 rules 3.2: corrected settlement and hindsight bound, refilled packs, and the rules-3.2 re-pilot 07:07 Rules 3.1/4.1/5.1: rebuilt packs and re-pilots, hindsight capture as every board's headline, the registered Qwen3-4B training run (part 4) 07:07 Rules 3.1/4.1/5.1: rebuilt packs and re-pilots, hindsight capture as every board's headline, the registered Qwen3-4B training run (part 3) 07:07 Rules 3.1/4.1/5.1: rebuilt packs and re-pilots, hindsight capture as every board's headline, the registered Qwen3-4B training run (part 2) 07:07 Rules 3.1/4.1/5.1: rebuilt packs and re-pilots, hindsight capture as every board's headline, the registered Qwen3-4B training run Oct 24 Yesterday 21:37 v4 Weather Front and v5 The Market Floor previews: tasks, references, the v4 forecast store, anonymised v5 market days and replay-verified GPT-6.1 Sol pilot episodes; training arm 2c (part 2) 21:36 v4 Weather Front and v5 The Market Floor previews: tasks, references, the v4 forecast store, anonymised v5 market days and replay-verified GPT-6.1 Sol pilot episodes; training arm 2c 16:26 Training: arm 2b, a 105-step GRPO continuation of arm 2 (0.668 sampled, 0.681 greedy) (part 2) 16:26 Training: arm 2b, a 105-step GRPO continuation of arm 2 (0.668 sampled, 0.681 greedy) 15:17 v2 and v3 previews: tasks, references and replay-verified GPT-6.1 Sol pilot episodes; training summary (part 3) 15:17 v2 and v3 previews: tasks, references and replay-verified GPT-6.1 Sol pilot episodes; training summary (part 2) 15:17 v2 and v3 previews: tasks, references and replay-verified GPT-6.1 Sol pilot episodes; training summary 12:16 Card: the hero image ships with the dataset; GitHub links point at main until the v0.1.0 tag exists 08:13 GridSimEnv v1: tasks, references and the GPT-6.1 Sol board episodes 07:34 initial commit 
