mlsanityData to 14 Oct 2026

SyRuP: Enhancing System-Prompt Following via Reward-Guided Prediction in LLM Decoding

arXiv 2607.23991 ↗First posted 27 Jul 2026cs.CL cs.AIcited by 0

On the map: LLM alignment / RLHF / preference optimisation → Reward modeling and RLHF

Abstract (arXiv)

Large Language Models (LLMs) are increasingly controlled through system prompts that specify roles, formats, and safety requirements. However, models follow these prompts only implicitly through in-context learning, which can be insufficient for complex or compositional prompts. Existing approaches often require model tuning or response-level reranking, limiting their practicality for lightweight inference-time control. We introduce SyRuP, a decoding-time framework for improving system-prompt adherence while keeping the base LM frozen. SyRuP trains a cross-attention reward head from system-prompt-conditioned preference pairs, treating the system prompt as a separate memory to produce token-level adherence scores. At inference, SyRuP reranks the base LM's top-k candidates by combining base logits with both the learned reward signal and a contrastive signal that captures system-induced logit shifts. Experiments on system-prompt following benchmarks show that SyRuP consistently outperforms prompting and decoding-time baselines with moderate inference overhead. These results suggest that explicit token-level guidance is an effective and practical mechanism for reliable system-prompt following.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reward modeling and RLHF, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 7 of its 50 references that are in the map. First posted on arXiv 27 Jul 2026.

Builds on

7 of 50 references.

References · 7

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2024Qwen2.5 Technical Reportcited by 4,942
  4. 2023Instruction-Following Evaluation for Large Language Modelscited by 1,228
  5. 2024DeepSeek-V3 Technical Reportcited by 811
  6. 2024The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructionscited by 493
  7. 2025A Closer Look at System Prompt Robustnesscited by 32

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026The Missing Piece in Pre-trained Model Evaluation: Reward-Guided Decoding Unlocks Task-Oriented Behavior Without Parameter Updates
  2. 2025Learning a Single Token to Replace Long System Prompts in LLMs
  3. 2025Cross-Lingual Prompt Steerability: Towards Accurate and Robust LLM Behavior across Languages
  4. 2026Robust Reward Modeling for Large Language Models via Causal Decomposition
  5. 2026PARTREP: Learning What to Repeat for Decoder-only LLMs
  6. 2025Sailing by the Stars: A Survey on Reward Models and Learning Strategies for Learning from Rewards
  7. 2024Self-Instructed Derived Prompt Generation Meets In-Context Learning: Unlocking New Potential of Black-Box LLMs
  8. 2026Maximizing Mutual Information Between Prompt and Response Improves LLM Performance With No Additional Data