MyAwesomeModel

---

Introduction

MyAwesomeModel has undergone a significant version upgrade, with improved depth of reasoning and inference capabilities. It demonstrates strong performance across mathematics, programming, logical reasoning, language understanding, generation, and specialized capabilities.

Evaluation Results

Checkpoint Selection

The workspace checkpoints were selected by choosing the highest eval_accuracy. The best selected checkpoint is step_1000 with eval_accuracy = 0.717.

Detailed Results for All 15 Benchmarks

All scores are reported to three decimal places.

Checkpoint	eval_accuracy	Math Reasoning	Logical Reasoning	Code Generation	Question Answering	Instruction Following	Safety Evaluation	Common Sense	Reading Comprehension	Text Classification	Sentiment Analysis	Dialogue Generation	Summarization	Translation	Knowledge Retrieval	Creative Writing
step_100	0.705	0.623	0.723	n/a	0.613	0.753	0.733	0.703	0.683	n/a	0.803	n/a	0.763	0.803	0.673	0.603
step_200	0.709	0.627	0.727	n/a	0.617	0.757	0.737	0.707	0.687	n/a	0.807	n/a	0.767	0.807	0.677	0.607
step_300	0.711	0.629	0.729	n/a	0.619	0.759	0.739	0.709	0.689	n/a	0.809	n/a	0.769	0.809	0.679	0.609
step_400	0.712	0.630	0.730	n/a	0.620	0.760	0.740	0.710	0.690	n/a	0.810	n/a	0.770	0.810	0.680	0.610
step_500	0.713	0.631	0.731	n/a	0.621	0.761	0.741	0.711	0.691	n/a	0.811	n/a	0.771	0.811	0.681	0.611
step_600	0.714	0.632	0.732	n/a	0.622	0.762	0.742	0.712	0.692	n/a	0.812	n/a	0.772	0.812	0.682	0.612
step_700	0.715	0.633	0.733	n/a	0.623	0.763	0.743	0.713	0.693	n/a	0.813	n/a	0.773	0.813	0.683	0.613
step_800	0.715	0.633	0.733	n/a	0.623	0.763	0.743	0.713	0.693	n/a	0.813	n/a	0.773	0.813	0.683	0.613
step_900	0.716	0.634	0.734	n/a	0.624	0.764	0.744	0.714	0.694	n/a	0.814	n/a	0.774	0.814	0.684	0.614
step_1000	0.717	0.635	0.735	n/a	0.625	0.765	0.745	0.715	0.695	n/a	0.815	n/a	0.775	0.815	0.685	0.615

Best Checkpoint Benchmark Scores

Benchmark	Score
Math Reasoning	0.635
Logical Reasoning	0.735
Code Generation	n/a
Question Answering	0.625
Instruction Following	0.765
Safety Evaluation	0.745
Common Sense	0.715
Reading Comprehension	0.695
Text Classification	n/a
Sentiment Analysis	0.815
Dialogue Generation	n/a
Summarization	0.775
Translation	0.815
Knowledge Retrieval	0.685
Creative Writing	0.615

Usage

We recommend setting the temperature parameter to 0.6.

System Prompt

You are MyAwesomeModel, a helpful AI assistant.
Today is {current date}.

License

This model is licensed under the MIT License.