2 days ago
-
Beam's effort dial next to the student's, the distillation recipe, and new scripts
3 days ago
-
Note the coding-data experiment under what did not help
4 days ago
-
Teacher comparison: AMC is a tie
-
Add Beam (teacher) scores on the same eval
-
Add budget-forcing results, EvalPlus coding results and the live ZeroGPU demo
8 days ago
-
v2: Beam reasoning-effort levels (low/medium/high), Accurate variant; v1 is tagged
-
Re-evaluate at a 64k-token limit: resume answers cut off at 32k (corrects AIME comparisons)
10 days ago
-
Video poster
-
Real-time M4 Max race video + raw timings; Mac speed note; race and eval scripts
-
Card: link published eval outputs; code: audit_eval.py, export_eval_data.py
-
Card: chart, IFEval, llama.cpp + Space links; code: IFEval, chart and Space builders
-
Model card: method, results at 1/2/4 loops, usage
-
Dory-Beam-4B weights, config, tokenizer, license and code
-
initial commit
