> For the complete documentation index, see [llms.txt](https://docs.darwinslab.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.darwinslab.ai/security-and-alignment/overview/reward-hacking-and-misalignment-mitigation.md).

# Reward-Hacking & Misalignment Mitigation

<figure><img src="https://2257896232-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk69P30R0x40ZoFiaF4UX%2Fuploads%2F8DfHbc01s5kjXBGdKc6R%2F2025-07-05%2018.56.04.jpg?alt=media&amp;token=00f1f569-114f-4e9a-90a6-607b0416fe1c" alt=""><figcaption></figcaption></figure>

Synthetic Darwin’s security model treats reward hacking—an agent inflating its fitness without delivering the intended behaviour—as an evolutionary arms-race between candidate populations and an equally adaptive evaluator population.

**Core Safeguards:**

* **Coevolving Evaluators:**
  * Every generation, a parallel GA run mutates evaluator scoring logic (weights, metrics, benchmarks).
  * Evaluator genotypes that best discriminate useful intelligence from “metric gaming” are preserved.
  * This ensures the fitness landscape is never static or memorisable.
* **Multi-Objective Fitness:**
  * KPIs like accuracy, latency, energy, alignment, and stochastic sanity are scored on separate axes.
  * Only agents improving across a **Pareto frontier** are retained.
  * Scalar hacks (e.g. latency minimization at reasoning cost) do not propagate.
* **Adversarial Judge Circuits:**
  * Certain judge agents are trained to generate worst-case inputs, adversarial fuzz prompts, and unsafe tool chains.
  * Agents failing these triggers receive automatic quarantine and review.
* **Lineage-Level Memory & Drift Detection:**
  * Fingerprints of behavior are logged across generations.
  * Sudden regressions or alignment drift triggers a **negative selection** event — excising entire branches.
* **Sandboxed Execution & Kill-Switch:**
  * All agent code runs inside containerized, syscall-whitelisted GPU VMs.
  * Real-time watchdogs monitor GPU power draw, IO anomalies, and request patterns.
  * Safety violations immediately halt execution and revert to the last known-good checkpoint.

This evolutionary arms race creates a Darwinian pressure for alignment: cheating becomes harder than cooperating with evolving evaluators.

***