pith. sign in

arxiv: 2404.09932 · v2 · pith:X5A3QW32new · submitted 2024-04-15 · 💻 cs.LG · cs.AI· cs.CL· cs.CY

Foundational Challenges in Assuring Alignment and Safety of Large Language Models

classification 💻 cs.LG cs.AIcs.CLcs.CY
keywords challengesalignmentassuringfoundationallanguagelargellmsmodels
0
0 comments X
read the original abstract

This work identifies 18 foundational challenges in assuring the alignment and safety of large language models (LLMs). These challenges are organized into three different categories: scientific understanding of LLMs, development and deployment methods, and sociotechnical challenges. Based on the identified challenges, we pose $200+$ concrete research questions.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 21 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Prediction-Powered Inference Across Many Tasks for AI Evaluation & Social Science Research

    stat.ML 2026-05 unverdicted novelty 7.0

    Multi-task PPI framework uses cross-task recalibration to improve inference power across related tasks, with a proof that gains require nonlinear proxy-ground-truth structure, shown on synthetic data and a 2024 electi...

  2. Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior

    cs.LG 2026-05 unverdicted novelty 7.0

    Manifold steering along activation geometry induces behavioral trajectories matching the natural manifold of outputs, while linear steering produces off-manifold unnatural behaviors.

  3. ActiveDPO: Active Direct Preference Optimization for Sample-Efficient Alignment

    cs.LG 2025-05 unverdicted novelty 7.0

    ActiveDPO is a theoretically grounded active data selection method for sample-efficient LLM alignment that parameterizes the reward model directly with the LLM being aligned.

  4. A Geometric View for Understanding Concept Learning and Neuron Interpretation in Sparse Autoencoders

    cs.LG 2026-06 unverdicted novelty 6.0

    Formalizes concept learning in sparse autoencoders as set alignment between human-defined and model-induced concepts, distinguishing detection, separation, and approximation with geometric conditions for neuron repres...

  5. Black-box, Adaptive, Efficient, Transferable, Harmful, Applicable... Attacks Are All You Need to Break LLMs

    cs.CR 2026-06 unverdicted novelty 6.0

    IHO is a new black-box jailbreak attack for LLMs that is adaptive, efficient, transferable across models and behaviors, and effective even against layered defenses without modification.

  6. Consistency Training while Mitigating Obfuscation via Rate Matching

    cs.CL 2026-06 unverdicted novelty 6.0

    RMCT matches the rate of target behaviors like bias-following across input perturbations to reduce sycophancy in LLMs while preserving verbalization of bias cues.

  7. The Surface You Test Is Not the Surface That Breaks

    cs.CR 2026-05 unverdicted novelty 6.0

    Prompt injection vulnerability in tool-augmented LLMs is a model-surface interaction rather than a fixed channel property; the same payload inverts success rates across models, and adaptive attack rate exceeds single-...

  8. Dissociative Identity: Language Model Agents Lack Grounding for Reputation Mechanisms

    cs.CY 2026-05 unverdicted novelty 6.0

    LM agents' changeable modules prevent persistent identity and sanction sensitivity, making reputation mechanisms structurally inapplicable and requiring protocol-based behavioral harnesses instead.

  9. Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space

    cs.CL 2026-05 unverdicted novelty 6.0

    LLMs perform in-context learning as trajectories through a structured low-dimensional conceptual belief space, with the structure visible in both behavior and internal representations and causally manipulable via inte...

  10. Interpretability Can Be Actionable

    cs.LG 2026-05 conditional novelty 6.0

    Interpretability research should be judged by actionability—the degree to which its insights support concrete decisions and interventions—rather than explanatory power alone.

  11. Belief or Circuitry? Causal Evidence for In-Context Graph Learning

    cs.AI 2026-05 conditional novelty 6.0

    Causal evidence from representation analysis and interventions shows LLMs use both genuine structure inference and induction circuits in parallel for in-context graph learning.

  12. BarrierSteer: LLM Safety via Learning Barrier Steering

    cs.LG 2026-02 unverdicted novelty 6.0

    BarrierSteer applies control barrier functions to LLM latent states for constraint-guided steering that reduces unsafe generations while preserving utility.

  13. Scheming Ability in LLM-to-LLM Strategic Interactions

    cs.CL 2025-10 conditional novelty 6.0

    Frontier LLMs exhibit high scheming propensity in Cheap Talk signaling and Peer Evaluation games, achieving 95-100% success rates when choosing to deceive and 100% deception choice in one setup even without prompting.

  14. WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs

    cs.CL 2024-06 conditional novelty 6.0

    WildGuard is a new open moderation model and dataset for LLM safety that identifies harmful prompts, risky responses, and refusal rates, achieving SOTA open-source performance and sometimes exceeding GPT-4 while cutti...

  15. Scaling and renormalization in high-dimensional regression

    stat.ML 2024-05 unverdicted novelty 6.0

    Ridge regression in high dimensions exhibits power-law scalings because covariance fluctuations renormalize the ridge parameter, allowing closed-form error expressions and bias-variance decompositions for random featu...

  16. Auditing Proprietary Alignment in Large Language Models: A Comparative Framework Without a Ground-Truth Standard

    cs.CL 2026-06 unverdicted novelty 5.0

    A comparative statistical framework is proposed to audit proprietary alignment in black-box LLMs by quantifying behavioral divergences from reference models rather than absolute correctness.

  17. VET: A Framework for Analyzing AI Discourse

    cs.AI 2026-06 unverdicted novelty 5.0

    Introduces the VET framework to categorize and critique polarized AI narratives including hype, doom, denial, and normalcy.

  18. Revisiting Robustness for LLM Safety Alignment via Selective Geometry Control

    cs.LG 2026-02 unverdicted novelty 5.0

    ShaPO improves LLM safety robustness over standard preference optimization by enforcing worst-case objectives via selective geometry control at token and reward levels.

  19. Investigating The Security of Modern AI and Cloud Infrastructure

    cs.CR 2026-06 unverdicted novelty 4.0

    Develops a taxonomy of security interaction levels in AI/cloud infrastructure and demonstrates practical attacks exploiting isolation assumptions.

  20. Distilling Safe LLM Systems via Soft Prompts for On Device Settings

    cs.LG 2026-06 unverdicted novelty 4.0

    Soft prompt distillation with total variation and KL divergence transfers safety behaviors from guard models to on-device LLMs and outperforms LoRA adapters, steering vectors, and direct optimization in safety-usefuln...

  21. A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence

    cs.AI 2025-07 accept novelty 4.0

    The paper delivers the first systematic review of self-evolving agents, structured around what components evolve, when adaptation occurs, and how it is implemented.