Update tokenizer, configuration and training reports

reports/run/metrics.jsonl

@@ -418,3 +418,5 @@
 {"step": 8360, "tokens": 4383047680, "loss": 2.8418451309204102, "tokens_per_second": 980239.3377897158, "lr": 0.0015, "gradient_norm": 0.07009512186050415, "benchmark": false}
 {"step": 8380, "tokens": 4393533440, "loss": 2.840165710449219, "tokens_per_second": 988014.5265202852, "lr": 0.0015, "gradient_norm": 0.07453550398349762, "benchmark": false}
 {"step": 8400, "tokens": 4404019200, "loss": 2.865531349182129, "tokens_per_second": 839928.994479502, "lr": 0.0015, "gradient_norm": 0.07409143447875977, "benchmark": false}
+{"step": 8420, "tokens": 4414504960, "loss": 2.824263000488281, "tokens_per_second": 976282.8477343371, "lr": 0.0015, "gradient_norm": 0.07757251709699631, "benchmark": false}
+{"step": 8440, "tokens": 4424990720, "loss": 2.8189924240112303, "tokens_per_second": 983180.1213848125, "lr": 0.0015, "gradient_norm": 0.07665544003248215, "benchmark": false}