AGENTIC-SYSTEMSApril 22, 20261 min read7 views

ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System

ARES is a novel framework designed to systematically discover and mitigate dual vulnerabilities in both LLMs and Reward Models, establishing a new paradigm for comprehensive RLHF safety alignment.

Computer Science > Artificial Intelligence

Title:ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System

Reinforcement Learning from Human Feedback (RLHF) is central to aligning Large Language Models (LLMs), yet it introduces a critical vulnerability: an imperfect Reward Model (RM) can become a single point of failure when it fails to penalize unsafe behaviors. While existing red-teaming approaches primarily target policy-level weaknesses, they overlook what we term systemic weaknesses cases where both the core LLM and the RM fail in tandem.

We present ARES, a framework that systematically discovers and mitigates such dual vulnerabilities. ARES employs a "Safety Mentor" that dynamically composes semantically coherent adversarial prompts by combining structured component types (topics, personas, tactics, goals) and generates corresponding malicious and safe responses. This dual-targeting approach exposes weaknesses in both the core LLM and the RM simultaneously. Using the vulnerabilities gained, ARES implements a two-stage repair process: first fine-tuning the RM to better detect harmful content, then leveraging the improved RM to optimize the core model. Experiments across multiple adversarial safety benchmarks demonstrate that ARES substantially enhances safety robustness while preserving model capabilities, establishing a new paradigm for comprehensive RLHF safety alignment.

Source: arXiv cs.AI Recent

More in this category

agentic-systems

FindStatBench: Evaluating Large Language Models on Combinatorial Code Synthesis

Researchers have introduced FindStatBench, a new execution benchmark containing 2,329 tasks designed to evaluate the combinatorial code synthesis capabilities of large language models.

agentic-systems

Probabilistic Concept-Aware Steering for Trustworthy LLM Inference

Researchers have introduced Probabilistic Concept-Aware Steering (PCS), an inference-time intervention framework for LLMs. PCS provides fine-grained, safety-oriented semantic steering while preserving original task competence.

agentic-systems

MUX: Continuous Reasoning via Multiplexed Tokens

Researchers introduced MUX, a novel latent reasoning method that compresses discrete text-based reasoning steps into continuous multiplexed tokens. By enabling lossless superposition, MUX significantly boosts LLM reasoning efficiency and speed across complex problem-solving tasks.

agentic-systems

S2T-RLHF: Hierarchical Credit Assignment for Stable Preference-Based RLHF

Researchers introduced S2T-RLHF, a hierarchical credit assignment framework that stabilizes preference-based RLHF training by decomposing response-level rewards into sentence and token levels.

agentic-systems

Semantic Cooperative Games for Contribution Attribution in LLM-Based Multi-Agent Systems

Researchers introduce Semantic Cooperative Games (SCG) and the SLIC algorithm to evaluate agent contributions in LLM-based multi-agent systems without re-running models, reducing computational costs by 93.3%.

agentic-systems

ProbSPARQL: Querying Knowledge Graphs with Multi-dimensional, Uncertain Numeric Data

Researchers have introduced ProbSPARQL, an upward-compatible SPARQL extension designed to query multi-dimensional and uncertain numeric measurement data within Knowledge Graphs, providing significant query performance gains for complex industrial applications.