fix: Duplicate predictions

[mhovd]

This PR fixes an issue where predictions were duplicated multiple times

This is something that should be covered by a test, some day...

[Copilot]

Pull Request Overview

This PR fixes a bug where predictions were being duplicated by changing the iteration logic to only use the first set of support point predictions instead of iterating over all predictions.

• Replaces nested iteration over all predictions with iteration over only the first support point predictions
• Adds explanatory comments to clarify the new logic
• Prevents duplicate prediction entries in the output

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[github-actions]

Bencher Report

Branch	fix-predictions
Testbed	pmcore-runner

Click to view all benchmark results

Benchmark	Latency	Benchmark Result seconds (s) (Result Δ%)	Upper Boundary seconds (s) (Limit %)
bimodal_ke_npag	📈 view plot 🚷 view threshold	34.29 s (+49.81%) Baseline: 22.89 s	49.46 s (69.32%)
bimodal_ke_npod	📈 view plot 🚷 view threshold	44.24 s (+85.33%) Baseline: 23.87 s	70.54 s (62.72%)
bimodal_ke_postprob	📈 view plot 🚷 view threshold	15.40 s (+85.14%) Baseline: 8.32 s	24.49 s (62.86%)

🐰 View full continuous benchmarking report in Bencher

[github-actions]

Bencher Report

Branch	fix-predictions
Testbed	rust-loki

Click to view all benchmark results

Benchmark	Latency	Benchmark Result seconds (s) (Result Δ%)	Upper Boundary seconds (s) (Limit %)
bimodal_ke_npag	📈 view plot 🚷 view threshold	34.75 s (+1.33%) Baseline: 34.29 s	48.85 s (71.14%)
bimodal_ke_npod	📈 view plot 🚷 view threshold	44.17 s (+0.17%) Baseline: 44.09 s	46.43 s (95.12%)
bimodal_ke_postprob	📈 view plot 🚷 view threshold	15.41 s (+0.36%) Baseline: 15.35 s	17.09 s (90.17%)

🐰 View full continuous benchmarking report in Bencher