Automatically jailbreaks LLMs with black-box access.
problem Generating harmful content from black-box LLMs.
method Automated method using an attacker LLM to refine prompts.
result Generates jailbreaks for over 80% of prompts.
On-device research index
arXiv research
A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
168,657 papers · 148 categories
Trend · papers per month
48 results for “LLM jailbreaking”
SmoothLLM defends LLMs from jailbreaking attacks by randomly perturbing inputs.
problem Adversaries can fool large language models into generating objectionable content.
method SmoothLLM randomly perturbs multiple copies of a prompt and aggregates predictions to detect adversarial inputs.
result SmoothLLM sets the state-of-the-art for robustness against various jailbreak attacks.
VERA uses variational inference to jailbreak LLMs without manual optimization.
problem Lack of principled objective for gradient-based optimization in jailbreaking LLMs.
method VERA casts black-box jailbreak prompting as a variational inference problem, training a small attacker LLM to approximate the target LLM's posterior over adversarial prompts.
result VERA achieves strong performance across various target LLMs, demonstrating the value of probabilistic inference for adversarial prompt generation.
Shorter adversarial prompts help protect LLMs from jailbreak attacks.
problem Protecting large language models from jailbreak attacks with long adversarial suffixes.
method Adversarial training on shorter adversarial suffixes to defend against longer adversarial suffixes.
result Aligning LLMs on shorter adversarial suffixes can effectively defend against jailbreak attacks with longer suffixes.
Understanding and Improving Continuous Adversarial Training for LLMs via In-context Learning Theorycs.LG
This paper improves continuous adversarial training for LLMs using in-context learning theory.
problem Efficiently defending large language models (LLMs) against jailbreak attacks.
method The paper presents a theoretical analysis of continuous adversarial training (CAT) for LLMs based on in-context learning (ICL) theory, proving a robust generalization bound and proposing an improved regularization term.
result The robust generalization bound explains why CAT can defend against jailbreak prompts and shows that LLM robustness is related to embedding matrix singular values.
RESTA defends LLMs against jailbreaking attacks by adding random noise to embeddings.
problem Vulnerability of LLMs to jailbreaking attacks that generate harmful outputs.
method Adds random noise to embedding vectors and aggregates during token generation.
result RESTA achieves superior robustness versus utility tradeoffs compared to baseline defenses.
Simple adaptive attacks can jailbreak state-of-the-art safety-aligned language models.
problem Vulnerability of safety-aligned language models to simple adaptive attacks.
method Designing adversarial prompt templates, random search on suffixes, and transfer/prefilling attacks.
result 100% attack success rate on various models including GPT-4, Vicuna, Mistral, and Claude.
COCA refactors training data to identify and erase unsafe concepts in LLMs.
problem Identifying and erasing unsafe concepts in Large Language Models (LLMs) for safety alignment.
method Concept Concentration (COCA) refactors training data with an explicit reasoning process to identify and erase unsafe concepts.
result COCA significantly reduces both in-distribution and out-of-distribution jailbreak success rates while maintaining strong performance on regular tasks.
High-dimensional models pose both safety benefits and risks.
problem Emergent problems in safety alignment due to high-dimensional representations.
method Detailed visualizations and lower-dimensional subspace projections.
result Dimensional reduction preserves safety alignment while avoiding linear jailbreaking.
Diffusion LLMs can efficiently generate harmful prompts for adversarial testing.
problem Generating harmful prompts for adversarial testing is resource-intensive and costly.
method Transformed adversarial prompt optimization into an efficient inference task using pretrained Diffusion LLMs.
result Only a few conditional samples are required to generate harmful prompts with high reward.
New framework assesses LLM security risks in BFSI.
problem Lack of domain-specific security evaluation for LLMs in BFSI.
method Risk-aware evaluation framework combining taxonomy, automated red-teaming, and ensemble judging.
result Higher decoding stochasticity and adaptive interaction lead to more severe disclosures.
Theoretical analysis shows LLMs can self-correct responses through in-context learning.
problem Understanding how large language models improve through self-correction.
method Theoretical analysis based on simplified alignment task, focusing on softmax attention, multi-head attention, and MLP blocks.
result LLMs can refine responses in an in-context way when given accurate self-examinations as rewards.
VERA-V uses variational inference to discover vulnerabilities in multimodal vision-language models.
problem Existing methods for jailbreaking vision-language models are brittle, limited, and focus on single attacks.
method VERA-V recasts jailbreak discovery as learning a joint posterior distribution over text-image prompts, using variational inference and three complementary strategies.
result VERA-V consistently outperforms state-of-the-art baselines, achieving up to 53.75% higher attack success rate.
ELS framework improves safety alignment by dynamically steering LLMs towards helpful responses.
problem Over-Refusal in Aligned Large Language Models
method Fine-tuning free framework using an Energy-Based Model (EBM) to dynamically steer LLMs during inference.
result Extensive experiments show a significant reduction in false refusals (from 57.3% to 82.6%) while maintaining safety performance.
This paper formalizes AI safety using hypothesis testing in GenAI.
problem Ensuring safety of generative AI tools that create realistic content.
method Formalization of computational safety through hypothesis testing and signal processing.
result Demonstrates how AI safety can be assessed quantitatively using mathematical frameworks.
New method accounts for hidden context in preference learning for RLHF models.
problem Incomplete data with hidden context affects RLHF model outcomes.
method Distributional Preference Learning (DPL) methods estimate hidden context distributions.
result DPL methods reduce RLHF vulnerabilities by accounting for hidden context.
Forecasting Credit Ratings: A Case Study where Traditional Methods Outperform Generative LLMsq-fin.RM
Traditional methods outperform LLMs in forecasting corporate credit ratings.
problem Forecasting corporate credit ratings using LLMs.
method Comparison of traditional methods (XGBoost) and LLMs (LLaMA) on credit rating forecasting.
result XGBoost outperforms LLMs in forecasting corporate credit ratings.
Unified framework to bridge human and LLM judgments.
problem Systematic discrepancies between human and LLM evaluations.
method Latent human preference score and linear transformations of covariates.
result Higher agreement with human ratings and exposure of systematic gaps.
LLM-Lasso uses LLMs to improve feature selection in Lasso regression.
problem Improving feature selection in Lasso regression with domain-specific knowledge.
method Combines LLMs with Lasso regularization to generate feature weights.
result Outperforms standard Lasso and feature selection baselines in biomedical studies.
THERMOMETER calibrates LLMs efficiently for diverse tasks.
problem Calibrating large language models is challenging due to computational and versatility issues.
method THERMOMETER learns an auxiliary model for calibrating a LLM using data from multiple tasks.
result THERMOMETER produces better-calibrated responses for new tasks.
Regurgitative training with synthetic data harms LLM performance.
problem The impact of training LLMs with synthetic data generated by other LLMs.
method Fine-tuning GPT-3.5 with synthetic and real data in a machine translation task.
result Regurgitative training significantly reduces LLM performance.
FLUID-LLM: Learning Computational Fluid Dynamics with Spatiotemporal-aware Large Language Modelscs.LG
FLUID-LLM uses LLMs to predict fluid dynamics with improved accuracy.
problem Leveraging LLMs for CFD due to their pattern recognition abilities but struggles with fluid dynamics complexities.
method Combines pre-trained LLMs with spatiotemporal-aware encoding to predict unsteady fluid dynamics.
result Significant performance improvements in CFD predictions across various datasets.
LLMs mimic human traders in finance, but not as much as expected.
problem Evaluating how LLMs behave in financial markets.
method Adapted experimental design with LLMs and human traders, analyzed in single and mixed model settings.
result LLMs tend to price assets near their fundamental value, but not as much as humans, and show less trading strategy variance.
Paper presents a probabilistic model to improve LLM cascade performance.
problem Complexity of LLM cascades and their interaction error rates.
method Probabilistic model for joint performance distribution of LLMs.
result Improves area under the error-cost curve by 4.3% on average for cascades with k≥3 models.
LLMs struggle to optimize hyperparameters efficiently, but hybrid methods can improve performance.
problem Optimizing hyperparameters of small language models using LLMs.
method Comparison of classical HPO algorithms and LLM-based methods, introducing Centaur hybrid approach.
result Hybrid Centaur approach achieves best results, outperforming classical and pure LLM methods.
Improves feature selection in high-dimensional data using LLM-generated weights.
problem Inaccurate LLM-generated weights degrade feature selection performance.
method Integrates LLM-generated weights into prior inclusion probabilities using LLM Sparsity Prior (LSP).
result Improves prediction accuracy and identifies clinically relevant features.
Telescope detects LLM generated text by measuring token repetition probability.
problem Distinguishing LLM generated text from human writing.
method Telescope Perplexity, evaluating token repetition probability.
result Telescope Perplexity enables effective zero-shot LLM detection.
BED-LLM uses Bayesian experimental design to improve LLMs' information gathering.
problem Improving LLMs' ability to gather information adaptively.
method Iteratively choosing questions to maximize expected information gain using a probabilistic model.
result BED-LLM achieves substantial performance gains compared to other adaptive design strategies.
New benchmarks show LLMs struggle with causal discovery.
problem Leveraging LLMs for causal discovery is unreliable due to dataset leakage.
method Developing science-grounded benchmarks and hybrid methods combining LLM predictions with statistical analysis.
result LLMs perform poorly on novel, real-world scientific studies compared to classical methods.
Tiny benchmarks reduce LLM evaluation costs by using fewer examples.
problem Expensive evaluation of LLMs with tens of thousands of examples.
method Developed evaluation tools and tiny versions of popular benchmarks.
result Accurately estimate LLM performance with just 100 curated examples.
SelfReflect measures LLM uncertainty by summarizing belief distribution.
problem LLMs lack transparency in expressing uncertainty.
method Developed SelfReflect metric to quantify uncertainty.
result Modern LLMs cannot reflect their uncertainty.
LLMs overestimate stock returns and are less accurate at predicting extreme outcomes.
problem Behavioral biases in LLMs' stock return forecasts.
method Comparison of LLM forecasts with crowd-sourced estimates and historical data.
result LLMs overestimate stock returns and are less accurate at predicting extreme outcomes.
Survey of LLMs in finance tasks, including adoption and performance.
problem Utilizing large language models in financial tasks.
method Review of current approaches, decision framework for adoption.
result Synthesizes state-of-the-art for LLMs in finance.
Statsformer validates and adapts LLM-derived semantic priors for improved supervised learning.
problem Unreliable semantic priors from LLMs can degrade supervised learning performance.
method Adapts LLM-derived feature scores into a family of learner-specific prior-injection mechanisms, calibrating their influence using out-of-fold validation.
result Improves prediction performance by adaptively downweighting unreliable LLM priors, ensuring a guardrailed statistical learning system.
StealthRank subtly boosts LLM rankings without detectable anomalies.
problem Adversarial manipulation of LLM-driven ranking systems.
method Energy-based optimization with Langevin dynamics for stealthy prompt generation.
result StealthRank outperforms existing methods in covertly boosting rankings.
No-knowledge alarms detect misaligned LLM judges without trusting them.
problem Monitoring LLM judges for consistency when ground truth is unknown.
method Formal logic and Linear Programming to detect disagreements.
result No false positives in detecting misaligned judges.
Debiasing LLMs by Fine-tuningq-fin.GN
Fine-tunes LLMs to correct bias in predictions.
problem LLMs exhibit bias in predictions from data.
method Supervised fine-tuning with Low-Rank Adaptation (LoRA).
result Fine-tuning corrects bias in both controlled and real-world settings.
AlphaPruning optimizes LLM pruning using HT-SR theory for better performance.
problem Improving pruning of large language models to reduce size without sacrificing performance.
method AlphaPruning uses HT-SR theory to allocate layerwise sparsity ratios more theoretically.
result AlphaPruning prunes LLaMA-7B to 80% sparsity with reasonable perplexity.
The paper introduces Relative Bias to quantify LLM bias systematically.
problem Quantifying bias in LLMs is challenging due to ambiguity and rapid model emergence.
method Relative Bias framework using Embedding Transformation and LLM-as-a-Judge methodologies.
result The two scoring methods show strong alignment, providing a systematic approach.
Survey examines LLMs in financial trading.
problem Using LLMs to outperform professional traders in finance.
method Comprehensive review of current research on LLMs in financial trading.
result LLMs can potentially outperform professional traders in backtesting.
Method constructs finance LLMs without instruction data using pretraining and model merging.
problem Developing domain-specific LLMs for finance is resource-intensive.
method Continual pretraining on financial data + model merging of instruction-tuned and domain-specific pretrained vectors.
result Successfully constructs instruction-tuned LLMs for finance without additional instruction data.
Bayesian approach quantifies uncertainty in LLM-based systems.
problem Uncertainty quantification in LLM-based systems, especially for high-stakes applications.
method Interpreting prompts as parameters in a Bayesian model, using MHLP for inference.
result Improvements in predictive accuracy and uncertainty quantification on various benchmarks.
MultiwayPAM clusters LLM-as-a-Judge scores to reveal evaluator bias.
problem Revealing bias in LLM evaluations of text quality.
method Tensor clustering with MultiwayPAM method.
result Medoids reveal cluster membership of questions, answerers, and evaluators.
LLMs can help explain credit risk models but not autonomously.
problem Leveraging LLMs for post-hoc explainability in credit risk models.
method Comparison of LLM outputs with SHAP and coefficient-based attributions on three LMs.
result LLMs reliably preserve feature-importance rankings but poorly align with autonomous explanations.
New attack reveals memorization patterns in pre-trained LLMs.
problem Determining if a data point was part of a pre-trained LLM's training set.
method Adapts MIA statistical tests to LLM's perplexity dynamics of subsequences.
result Significantly outperforms prior approaches in membership inference attacks.
Study evaluates how much knowledge LLMs have by comparing their prediction accuracy to flexible models.
problem Evaluating the predictive power of LLMs without access to their training data.
method Equivalent sample size measure, comparing LLM's prediction error to flexible models trained on varying amounts of domain-specific data.
result LLMs encode varying amounts of predictive information across different economic variables.
This paper uses LLMs and cycle consistency for better machine translation evaluation.
problem Evaluating translation quality and LLM capabilities without ground truth.
method Generate translation candidates, back-translate, and evaluate cycle consistency.
result Larger LLMs or more inference passes improve cycle consistency.
LOLA uses LLMs to optimize content delivery, outperforming traditional methods.
problem Identifying the most engaging headlines for user engagement.
method LOLA integrates LLMs with adaptive experimentation to optimize content delivery.
result LOLA outperforms traditional methods in optimizing user engagement.