Challenge: Existing methods for enhancing harmlessness and helpfulness of large language models (LLMs) involve complex and resource-intensive training processes.
Approach: They propose a method that decouples harmlessness from helpfulness during inference phase.
Outcome: The proposed method significantly reduces the attack success rate (ASR) of harmful instructions and jailbreak instructions while maintaining almost unchanged performance in downstream tasks.

Similar Papers

PURE: Aligning LLM via Pluggable Query Reformulation for Enhanced Helpfulness (2024.findings-emnlp)

Copied to clipboard

Challenge: Large language models (LLMs) depend on vast amounts of text data sourced from the Internet for their training.
Approach: They propose a new alignment paradigm that reformulates risky queries into highly relevant yet harmless ones before feeding them into LLMs.
Outcome: The proposed approach eliminates the high costs of training base LLMs and achieves a promising balance of harmlessness and helpfulness.
Adaptive Helpfulness–Harmlessness Alignment with Preference Vectors (2026.eacl-long)

Copied to clipboard

Challenge: Existing approaches to balancing helpfulness and harmlessness suffer from performance conflicts, limited controllability, and poor extendability.
Approach: They propose a framework that allows users to control their own preferences and dynamically merge them at test time.
Outcome: The proposed framework improves helpfulness without conservatism and smooth control over preference trade-offs.
Multi-Attribute Steering of Language Models via Targeted Intervention (2025.acl-long)

Copied to clipboard

Challenge: Existing approaches for steering large language models fail to scale to multi-attribute settings with conflicts, such as enhancing helpfulness while also reducing toxicity.
Approach: They propose a steering framework for selective token-level intervention across multiple attributes that enforcing sparsity and orthogonality among vectors for different attributes.
Outcome: The proposed framework outperforms existing ITI and parameter-efficient fine-tuning approaches across question answering tasks and generative tasks.
Safety Arithmetic: A Framework for Test-time Safety Alignment of Language Models by Steering Parameters and Activations (2024.emnlp-main)

Copied to clipboard

Challenge: Current alignment methods struggle with dynamic user intentions and complex objectives, making models vulnerable to harmful content.
Approach: They propose a training-free framework that enhances LLM safety across different scenarios.
Outcome: The proposed framework significantly improves safety measures, reduces over-safety, and maintains model utility, outperforming existing methods in ensuring safe content generation.
Calibrating Inference Time Alignment with Sequence-level Risk Accumulation (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to decode large language models (LLMs) often over-reject benign information, limiting their generalizability in real-world scenarios where harmful and benign information coexist.
Approach: They propose a framework to regulate decoding alignments for Large Language Models (LLMs) they employ a reward-guided branch decoding paradigm to incorporate safety awareness during generation.
Outcome: The proposed framework achieves superior performance on four attack benchmarks and two neutral datasets.
To Intervene or Not: Guiding Inference-time Alignment with Probabilistic Model Blending (2026.findings-acl)

Copied to clipboard

Challenge: Existing approaches to inference-time alignment are expensive and only offer guidances during output generation.
Approach: They propose an inference-time alignment framework that shifts from binary decisions to creating hybrid distributions integrating both models’ knowledge.
Outcome: The proposed framework reduces the number of inference-time alignment interventions and improves performance on challenging model pairs.
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training (2026.findings-acl)

Copied to clipboard

Challenge: Misaligned large language models can magnify harm by exploiting them to undermine safety . et al., 2022b; Bai e.t., 2023): misalignment, realignment and model-specific resistance are important .
Approach: They evaluate four methods to identify a mechanism asymmetry between attack and defense . they find that ORPO is most effective for misalignment, but DPO excels in realignment .
Outcome: The proposed methods show a mechanism asymmetry between attack and defense . the proposed methods excel in realignment, but at the expense of model utility .
When the Model Said ‘No Comment’, We Knew Helpfulness Was Dead, Honesty Was Alive, and Safety Was Terrified (2026.eacl-long)

Copied to clipboard

Challenge: Existing work uses SFT and MoE to align Large Language Models, but these work face challenges in multi-objective settings.
Approach: They propose a framework that uses prompt-injected fine-tuning to extract axis-specific task features . it deploys a MoCaE module that calibrates expert routing using fractal and natural geometry .
Outcome: The proposed framework achieves significant gains on Alpaca, BeaverTails, TruthfulQA and TruthfulQ with +171.5% win rate and +110.1% truthfulness-informativeness.
DIESEL: A Lightweight Inference-Time Safety Enhancement for Language Models (2025.findings-acl)

Copied to clipboard

Challenge: Large language models generate outputs that are not aligned with human values, such as toxic content, malicious use cases, and vulnerabilities to adversarial jailbreak attacks.
Approach: They propose a lightweight inference-guidance technique that can be seamlessly integrated into any autoregressive LLM to semantically filter undesirable content during generation.
Outcome: The proposed technique can be integrated into any autoregressive LLM to semantically filter undesirable content during generation.
Beyond Reactive Safety: Risk-Aware LLM Alignment via Long-Horizon Simulation (2025.findings-acl)

Copied to clipboard

Challenge: Existing alignment methods focus on reactive feedback, where immediate human perception is leveraged to judge sampled model responses as preference data for post-training.
Approach: They propose a proof-of-concept framework that projects how model-generated advice could propagate through societal systems on a macroscopic scale over time, enabling more robust alignment.
Outcome: The proposed framework achieves 20% improvement on existing safety benchmarks and an average win rate exceeding 70% against strong baselines.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations