On the 100 English-instructed tool-use task cells with at most 8 steps and 2048 generated tokens per step, Bielik-PL-11B-v3.0-Instruct generates 806.6 tokens per trajectory on average and Bielik-11B-v3.0-Instruct 1011.6, a ratio of 0.7974 (95% bootstrap interval 0.5017 to 1.0846).

Relation: Metric ratio with an interval

metric

Mean generated tokens per trajectoryQuantity compared.

benchmark

Agentic task suiteTasks measured.

language

EnglishInstruction language of the cells.

setting

ReAct tool loop with a persistent Python sandbox: at most 8 steps, 2048 generated tokens per step, a FINAL answer also ends a trajectory in a turn with code once earlier code has executed; write-and-debug solutions rebuilt from executed code when no solution.py was written; MLX bf16, greedy decodingConditions of the measurement.

subject

Bielik-PL-11B-v3.0-InstructModel after the tokenizer change.

subject value

806.6 tokensMean for the subject.

comparator

Bielik-11B-v3.0-InstructModel before the tokenizer change.

comparator value

1011.6 tokensMean for the comparator.

value

0.7974 ratioSubject mean divided by comparator mean.

interval low

0.5017 ratioLower bound of the 95% bootstrap interval.

interval high

1.0846 ratioUpper bound of the 95% bootstrap interval.

Method and evaluation protocol

Divide the mean generated tokens per English-instructed trajectory of the APT4 model by the original's and bootstrap the ratio over paired cells (2000 resamples, seed 20260707).

Dataset

100 synthetic task instances from 20 parametric templates, each with English and Polish instructions.Version: tasks manifest sha256 66029e45c11f07613d698a67498b2c2c29faee4f0acc30bcf667bc3af6306f50 · Access: public

Reported results

Ratio 0.7974 [0.5017, 1.0846]; bootstrap p for a ratio of 1: 0.164, Holm-adjusted 0.328; the original model's turns contain a <think> block in 55 of 100 English cells.

Uncertainty and replication

95% percentile bootstrap interval over paired cells, 2000 resamples.

Evidence references

analysis.json · publichttps://github.com/stw2/tokenizer-science-tax/blob/76ed9b8efec6102935361df4b013764d057e4363/experiments/E10-agentic-compounding/results/analysis.jsongraded_summary.json · publichttps://github.com/stw2/tokenizer-science-tax/blob/76ed9b8efec6102935361df4b013764d057e4363/experiments/E10-agentic-compounding/results/graded_summary.jsonmetrics.json · publichttps://github.com/stw2/tokenizer-science-tax/blob/76ed9b8efec6102935361df4b013764d057e4363/experiments/E10-agentic-compounding/results/metrics.json

Limitations

H3 as designed compares prompt-side tokens; the analysis computes generated tokens. Single greedy run per cell on synthetic tasks; the two models differ by tokenizer, continued pretraining and post-training at once.