# Can Training Logs Make Model Comparisons More Precise?

arXiv: https://arxiv.org/abs/2608.02705 · HTML: https://mlsanity.com/paper/2608.02705/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 23 references that are in the map.
- First posted on arXiv 3 Aug 2026.

## Builds on (most cited first, 8 of 8)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Deep Reinforcement Learning that Matters](https://mlsanity.com/paper/1709.06560/) (2017)
- [Are GANs Created Equal? A Large-Scale Study](https://mlsanity.com/paper/1711.10337/) (2017)
- [Accounting for Variance in Machine Learning Benchmarks](https://mlsanity.com/paper/2103.03098/) (2021)
- [The Benchmark Lottery](https://mlsanity.com/paper/2107.07002/) (2021)
- [The MultiBERTs: BERT Reproductions for Robustness Analysis](https://mlsanity.com/paper/2106.16163/) (2021)
- [Torch.manual_seed(3407) is all you need: On the influence of random seeds in deep learning architectures for computer vision](https://mlsanity.com/paper/2109.08203/) (2021)

## Abstract (arXiv)

Comparing stochastically trained models requires estimating both a performance difference and its uncertainty from repeated runs. We study whether training logs from those same runs can make such comparisons more precise. Because training-log covariates are produced during training rather than measured before it, we use arm-specific covariate adjustment: each model is adjusted only with statistics from its own runs, and the raw mean difference remains the reported effect. In a vision study spanning three architectures and three datasets, simple adjustments based on early training logs often reduce uncertainty in model comparisons. The main limitation is covariate selection. Broadly searching the log pool for the most correlated statistic often adds more noise than it removes, even when useful statistics exist in hindsight. Training logs therefore appear useful for more precise model comparisons, but only when the adjustment avoids large selection noise.