[dev/newton] Tighten reward thresholds for Humanoid/G1/H1 training tests.

[adenzler-nvidia]

Description

I ran these tests locally and the reward thresholds can be tightened a bit. I would like to have them tighter because we will start running the training tests in the newton CI as a forward-compatibility check.

Type of change

• Test fix

Checklist

• I have read and understood the contribution guidelines
• I have run the pre-commit checks with ./isaaclab.sh --format
• [~] I have made corresponding changes to the documentation
• My changes generate no new warnings
• [~] I have added tests that prove my fix is effective or that my feature works
• [~] I have updated the changelog and the corresponding version in the extension's config/extension.toml file
• [] I have added my name to the CONTRIBUTORS.md or my name already exists there

[adenzler-nvidia]

@AntoineRichard are the CI failures expected here?

Anything else that needs to be done to get this merged?