{
  "annotation_id": "ff7092e1-601a-4c7b-a84d-e4393edfcfec",
  "date_created": "2026-02-17T05:53:08.236000Z",
  "date_modified": "2026-02-17T05:53:08.236000Z",
  "file_hash": "f4582ca37bf68ea5de24c81004c96a9fd3e3e57225f41654368f7ecbed3034f4",
  "private": false,
  "record": {
    "abstract": "Large language models excel at complex instructions yet struggle to deviate from their helpful assistant persona, as post-training instills strong priors that resist conflicting instructions. We introduce system prompt strength, a training-free method that treats prompt adherence as a continuous control. By contrasting logits from target and default system prompts, we isolate and amplify the behavioral signal unique to the target persona by a scalar factor alpha. Across five diverse benchmarks spanning constraint satisfaction, behavioral control, pluralistic alignment, capability modulation, and stylistic control, our method yields substantial improvements: up to +8.5 strict accuracy on IFEval, +45pp refusal rate on OffTopicEval, and +13% steerability on Prompt-Steering. Our approach enables practitioners to modulate system prompt strength, providing dynamic control over model behavior without retraining.",
    "arxiv_id": "2601.06403",
    "authors": [
      "Yijiang River Dong",
      "Tiancheng Hu",
      "Zheng Hui",
      "Nigel Collier"
    ],
    "categories": [
      "cs.CL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "title": "Steer Model beyond Assistant: Controlling System Prompt Strength via Contrastive Decoding",
    "url": "https://arxiv.org/abs/2601.06403",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "25f043ab-5b7a-47dc-8dd7-afa0c038a2fb",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}