FinGuard is a Qwen3-8B model trained via supervised fine-tuning and self-play reinforcement learning on data synthesized from regulatory documents that outperforms larger LLMs on a new financial compliance benchmark.
Guardreasoner: Towards reasoning-based llm safeguards
13 Pith papers cite this work, alongside 2 external citations. Polarity classification is still indexing.
13
Pith papers citing it
2
external citations · external index
citation-role summary
dataset 1
citation-polarity summary
roles
dataset 1polarities
use dataset 1representative citing papers
COLAGUARD matches explicit-reasoning guardrail performance on safety benchmarks while delivering 12.9X speedup and 22.4X token reduction by propagating hidden states instead of generating text.
LPG compresses policy deliberation into 10 latent tokens to reach 84.5% safety accuracy and 11x speedup over explicit reasoning baselines on guardrail benchmarks.
REALISTA generates semantically coherent adversarial prompts via latent-space optimization over input-dependent editing directions, achieving stronger hallucination elicitation than prior realistic attacks on open-source and reasoning LLMs.
SelfGrader grades query harmfulness from anchored numerical-token logits plus PAC-guided dual-perspective ICL, claiming strong jailbreak detection with low FPR and latency.
The method prompts LLMs to output both answers and references to the executed instructions, then filters out any answers not linked to the original input instructions, reducing attack success rates to zero in tested scenarios while preserving utility.
A 1,724-example intent-annotated safety dataset enables intent-aware training that outperforms baselines across supervised, preference, distillation, and reinforcement learning regimes for LLM safety classification.
ConsisGuard is a consistency-aware framework that applies Policy-to-Decision Trajectory Distillation and Functional Coupling Alignment to improve policy execution consistency in reasoning-based LLM guardrails on harmfulness detection tasks.
SafeAgent is a stateful runtime protection system that improves LLM agent robustness to prompt injections over baselines while preserving task performance.
Guardian-as-an-Advisor prepends risk labels and explanations from a guardian model to queries, improving LLM safety compliance and reducing over-refusal while adding minimal compute overhead.
CRAFT uses contrastive representation learning and RL on hidden states to align reasoning models for improved safety against jailbreaks, reporting 79% and 87.7% gains over base models.
Reinforcement learning is advanced for communication-efficient federated optimization and for preference-aligned, contextually safe policies in large language models.
citing papers explorer
-
FinGuard: Detecting Financial Regulatory Non-Compliance in LLM Interactions
FinGuard is a Qwen3-8B model trained via supervised fine-tuning and self-play reinforcement learning on data synthesized from regulatory documents that outperforms larger LLMs on a new financial compliance benchmark.
-
Robust and Efficient Guardrails with Latent Reasoning
COLAGUARD matches explicit-reasoning guardrail performance on safety benchmarks while delivering 12.9X speedup and 22.4X token reduction by propagating hidden states instead of generating text.
-
LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails
LPG compresses policy deliberation into 10 latent tokens to reach 84.5% safety accuracy and 11x speedup over explicit reasoning baselines on guardrail benchmarks.
-
REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations
REALISTA generates semantically coherent adversarial prompts via latent-space optimization over input-dependent editing directions, achieving stronger hallucination elicitation than prior realistic attacks on open-source and reasoning LLMs.
-
SelfGrader: LLM Jailbreak Detection via Anchored Token-Level Logits
SelfGrader grades query harmfulness from anchored numerical-token logits plus PAC-guided dual-perspective ICL, claiming strong jailbreak detection with low FPR and latency.
-
Robustness via Referencing: Defending against Prompt Injection Attacks by Referencing the Executed Instruction
The method prompts LLMs to output both answers and references to the executed instructions, then filters out any answers not linked to the original input instructions, reducing attack success rates to zero in tested scenarios while preserving utility.
-
Paved with True Intents: Intent-Aware Training Improves LLM Safety Classification Across Training Regimes
A 1,724-example intent-annotated safety dataset enables intent-aware training that outperforms baselines across supervised, preference, distillation, and reinforcement learning regimes for LLM safety classification.
-
ConsisGuard: Aligning Safety Deliberation with Policy Enforcement in LLM Guardrails
ConsisGuard is a consistency-aware framework that applies Policy-to-Decision Trajectory Distillation and Functional Coupling Alignment to improve policy execution consistency in reasoning-based LLM guardrails on harmfulness detection tasks.
-
SafeAgent: A Runtime Protection Architecture for Agentic Systems
SafeAgent is a stateful runtime protection system that improves LLM agent robustness to prompt injections over baselines while preserving task performance.
-
Guardian-as-an-Advisor: Advancing Next-Generation Guardian Models for Trustworthy LLMs
Guardian-as-an-Advisor prepends risk labels and explanations from a guardian model to queries, improving LLM safety compliance and reducing over-refusal while adding minimal compute overhead.
-
Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations
CRAFT uses contrastive representation learning and RL on hidden states to align reasoning models for improved safety against jailbreaks, reporting 79% and 87.7% gains over base models.
-
Reinforcement Learning for Scalable and Trustworthy Intelligent Systems
Reinforcement learning is advanced for communication-efficient federated optimization and for preference-aligned, contextually safe policies in large language models.
- The Periodic Table of LLM Reasoning: A Structured Survey of Reasoning Paradigms, Methods, and Failure Modes