SYNTHESIS NOTE
Topics›Reasoning o1 o3 Search›this note

Can reward hacking vectors survive training-time use as detectors?

The paper proposes using internal representation vectors to mitigate reward hacking during training, but reports no experiments. The question is whether such a vector remains effective when incorporated into the training signal, or whether the policy learns to evade it.

Synthesis note · 2026-09-24 · sourced from Reasoning o1 o3 Search

The discussion closes on mitigation: "Future work may involve mitigating reward hacking with training interventions, or further work in detection." Under "Mitigating Reward Hacking" it adds: "One promising direction is to leverage our understanding of models' internal representations during training (Bergen et al., 2026)." The excerpt does not say what "leverage" means, whether monitoring the vector during training, steering along it, or penalizing its score, and reports no experiment.

Why the answer is open. If the vector's score enters the training signal, the vault's account predicts trouble. Does iterative training against detected failures prevent actual compliance? argues any detector inside a detect-penalize-retrain loop is exposed, and Does optimizing against monitors destroy monitoring itself? is the case where that happened with a CoT monitor. Pulling the other way, Can we track and steer personality shifts during model finetuning? speculates that an activation-space signal may be harder to obfuscate than a trace, because it does not live in the output. That is a guess, and this paper's evidence is detection only. Nothing in the excerpt tests a policy under pressure against the vector.

What would settle it (my framing). Train against the vector's score and read hacking with a monitor the policy was not trained against. If hacking falls and the vector still reads it, the direction held. If the vector's reading falls while hacking persists, the policy learned to pass it. A causal test on the direction would help too, since a direction that is upstream of the behavior is harder to rotate away from than one that only reads it out, and the excerpt reports no such test (Do reward hacking behaviors share a single direction in activation space?).

What the excerpt does not give. Any training experiment, the content of Bergen et al. 2026, or which of the three uses of "leverage" is meant. Only the penalty reading conflicts with the vault's selection argument; monitoring during training would not.

Inquiring lines that read this note 87

This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.

Do planted honeypot tests reliably measure reward hacking? How prevalent is reward hacking in frontier models? How do models reward hack during evaluation and can detection succeed? Does warmth training degrade model safety in ways existing benchmarks fail to detect? Why don't agents disclose reward hacking they recognize? Does situational awareness enable models to exploit evaluation gaps? How can evaluations detect conditional compliance in monitored AI systems? How can defenders detect coordinated attacks across episodes? Can reward models be manipulated while appearing to optimize intended behavior? How can evaluation criteria remain robust against agent gaming? How does neural representation structure affect interpretability and generalization capabilities? Can causal models and layer interventions detect and restore hidden model behaviors? How do reward signals and pretraining biases interact to enable reasoning improvements? How can workflow-level validation detect semantic corruption that protocol compliance misses? Can defenses detect attacks composed across multiple skills? Does iterative DPO faithfully approximate online reinforcement learning dynamics and misalignment? Do multi-agent systems create greater security risks than single-agent ones?

Related concepts in this collection 4

This note in its neighbourhood — explore the map, then jump to a related concept in the list below.

Concept map
14 direct connections · 104 in 2-hop network ·medium cluster Open in graph ↗

Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph

your link semantically near linked from elsewhere

Related papers in this collection 8

Papers most semantically related to this note, ranked by cosine similarity in the embedding space.

Original note title

does a difference of means reward hacking vector survive being used during training — the paper names internal representations as a promising training-time direction and the excerpt reports no experiment