Commercial AI chatbots reach over 90% multiple-choice accuracy on recent news facts but lose 11-17% in free response and drop to 19-70% on subtle false-premise questions, with retrieval failures causing most errors and clear Anglophone bias.
Mixed citations
Title resolution pending
Mixed citation behavior. Most common role is background (50%).
40
Pith papers citing it
460
external citations · Crossref
Background
50% of classified citations
citation-role summary
background 4
dataset 1
method 1
citation-polarity summary
representative citing papers
LLMs routinely produce unsupported causal stories for personal sensing anomalies, and richer evidence or constrained prompts do not reliably eliminate this epistemic overreach.
A new paired-prompt protocol reveals alignment-pipeline-specific heterogeneity in how open-weight LLMs respond to evaluation versus deployment framings.
SENECA uses a novel self-consistent missing mass calculation to improve discrete entropy estimates in small-sample regimes and outperforms alternatives in numerical tests.
VLMs as judges exhibit informativeness bias by favoring detailed but image-inconsistent answers; BIRCH mitigates it by first correcting answers against the image, reducing bias up to 17% and improving performance up to 9.8%.
Alignment of vision-language models with human V1-V3 early visual cortex negatively predicts resistance to sycophantic gaslighting attacks.
Agreeableness in AI personas reliably predicts sycophantic behavior in 9 of 13 tested language models.
CapTrack shows post-training causes drift beyond facts, with instruction fine-tuning producing stronger behavioral changes than preference optimization across model families.
Norm-Anchor Scaling breaks the norm-feedback loop in sequential LLM editing by anchoring value vectors to original norms, improving long-run performance by 72.2% and extending the editing horizon over 4x.
LLM unlearning is reframed as inadvertently installing backdoor triggers on forget-tokens; Random Noise Augmentation is introduced as a defense that improves robustness with theoretical guarantees.
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
Chain-of-Thought reasoning in LLMs is often unfaithful, with models relying on it variably by task and less so as models scale larger.
Gradient-informed placement of LoRA parameters recovers full performance under GRPO while random placement does not, due to differences in gradient rank and stability across training regimes.
Under semantic underdetermination, LLMs cannot always guarantee strong correctness, strict non-bias, and high utility at once.
PRISM supplies a geometric upper bound on LLM variant risk that splits drift into scale, shape, and head axes and doubles as a differentiable regularizer against forgetting.
Symmetric spectral diagnostics on attention are structurally blind to flow direction, with asymmetry G as the sole control parameter, yielding a two-axis test that distinguishes bottleneck versus diffuse hallucination modes with opposite polarity.
Neuron-level inference-time intervention reduces multiple biases in reward models, enabling 2B and 7B models to match 70B performance on LLM alignment benchmarks without trade-offs.
LCF detects multiple LLM runtime threats by computing aggregated diagonal Mahalanobis distances on layer-wise hidden-state differences, calibrated on clean examples, achieving high detection rates with low overhead across several model architectures.
Architecture and training determine whether transformers retain a readable internal signal that lets activation monitors catch errors missed by output confidence.
ReFACT benchmark reveals LLMs show a persistent salient distractor failure mode where 61% of incorrect error span predictions are semantically unrelated to actual errors, persisting across model sizes, and comparative judgment yields lower F1 than independent detection.
Adapts multi-layer token-level Mahalanobis distance with supervised linear regression to yield improved uncertainty scores for LLM truthfulness tasks.
A regression model using attention features and recurrent uncertainty scores improves selective generation in LLMs over unsupervised and supervised baselines on ten datasets and three models.
Empirical analysis shows scaling inference compute via strategies like tree search can be more efficient than scaling model parameters, with 7B models plus novel search outperforming 34B models.
Sycophancy is prevalent in state-of-the-art AI assistants and is likely driven in part by human preferences that favor agreement over truthfulness.
citing papers explorer
-
Evaluating Commercial AI Chatbots as News Intermediaries
Commercial AI chatbots reach over 90% multiple-choice accuracy on recent news facts but lose 11-17% in free response and drop to 19-70% on subtle false-premise questions, with retrieval failures causing most errors and clear Anglophone bias.
-
Causal Stories from Sensor Traces: Auditing Epistemic Overreach in LLM-Generated Personal Sensing Explanations
LLMs routinely produce unsupported causal stories for personal sensing anomalies, and richer evidence or constrained prompts do not reliably eliminate this epistemic overreach.
-
Measuring Evaluation-Context Divergence in Open-Weight LLMs: A Paired-Prompt Protocol with Pilot Evidence of Alignment-Pipeline-Specific Heterogeneity
A new paired-prompt protocol reveals alignment-pipeline-specific heterogeneity in how open-weight LLMs respond to evaluation versus deployment framings.
-
SENECA: Small-Sample Discrete Entropy Estimation via Self-Consistent Missing Mass
SENECA uses a novel self-consistent missing mass calculation to improve discrete entropy estimates in small-sample regimes and outperforms alternatives in numerical tests.
-
When Vision-Language Models Judge Without Seeing: Exposing Informativeness Bias
VLMs as judges exhibit informativeness bias by favoring detailed but image-inconsistent answers; BIRCH mitigates it by first correcting answers against the image, reducing bias up to 17% and improving performance up to 9.8%.
-
Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic Manipulation
Alignment of vision-language models with human V1-V3 early visual cortex negatively predicts resistance to sycophantic gaslighting attacks.
-
Too Nice to Tell the Truth: Quantifying Agreeableness-Driven Sycophancy in Role-Playing Language Models
Agreeableness in AI personas reliably predicts sycophantic behavior in 9 of 13 tested language models.
-
CapTrack: Multifaceted Evaluation of Forgetting in LLM Post-Training
CapTrack shows post-training causes drift beyond facts, with instruction fine-tuning producing stronger behavioral changes than preference optimization across model families.
-
Norm Anchors Make Model Edits Last
Norm-Anchor Scaling breaks the norm-feedback loop in sequential LLM editing by anchoring value vectors to original norms, improving long-run performance by 72.2% and extending the editing horizon over 4x.
-
Improving LLM Unlearning Robustness via Random Perturbations
LLM unlearning is reframed as inadvertently installing backdoor triggers on forget-tokens; Random Noise Augmentation is introduced as a defense that improves robustness with theoretical guarantees.
-
GAIA: a benchmark for General AI Assistants
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
-
Measuring Faithfulness in Chain-of-Thought Reasoning
Chain-of-Thought reasoning in LLMs is often unfaithful, with models relying on it variably by task and less so as models scale larger.
-
Not How Many, But Which: Parameter Placement in Low-Rank Adaptation
Gradient-informed placement of LoRA parameters recovers full performance under GRPO while random placement does not, due to differences in gradient rank and stability across training regimes.
-
A CAP-like Trilemma for Large Language Models: Correctness, Non-bias, and Utility under Semantic Underdetermination
Under semantic underdetermination, LLMs cannot always guarantee strong correctness, strict non-bias, and high utility at once.
-
PRISM: A Geometric Risk Bound that Decomposes Drift into Scale, Shape, and Head
PRISM supplies a geometric upper bound on LLM variant risk that splits drift into scale, shape, and head axes and doubles as a differentiable regularizer against forgetting.
-
Self-Attention as Transport: Limits of Symmetric Spectral Diagnostics
Symmetric spectral diagnostics on attention are structurally blind to flow direction, with asymmetry G as the sole control parameter, yielding a two-axis test that distinguishes bottleneck versus diffuse hallucination modes with opposite polarity.
-
Debiasing Reward Models via Causally Motivated Inference-Time Intervention
Neuron-level inference-time intervention reduces multiple biases in reward models, enabling 2B and 7B models to match 70B performance on LLM alignment benchmarks without trade-offs.
-
Layerwise Convergence Fingerprints for Runtime Misbehavior Detection in Large Language Models
LCF detects multiple LLM runtime threats by computing aggregated diagonal Mahalanobis distances on layer-wise hidden-state differences, calibrated on clean examples, achieving high detection rates with low overhead across several model architectures.
-
Architecture Determines Observability of Transformers
Architecture and training determine whether transformers retain a readable internal signal that lets activation monitors catch errors missed by output confidence.
-
ReFACT: A Benchmark for Scientific Confabulation Detection with Positional Error Annotations
ReFACT benchmark reveals LLMs show a persistent salient distractor failure mode where 61% of incorrect error span predictions are semantically unrelated to actual errors, persisting across model sizes, and comparative judgment yields lower F1 than independent detection.
-
Token-Level Density-Based Uncertainty Quantification Methods for Eliciting Truthfulness of Large Language Models
Adapts multi-layer token-level Mahalanobis distance with supervised linear regression to yield improved uncertainty scores for LLM truthfulness tasks.
-
Unconditional Truthfulness: Learning Unconditional Uncertainty of Large Language Models
A regression model using attention features and recurrent uncertainty scores improves selective generation in LLMs over unsupervised and supervised baselines on ten datasets and three models.
-
Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models
Empirical analysis shows scaling inference compute via strategies like tree search can be more efficient than scaling model parameters, with 7B models plus novel search outperforming 34B models.
-
Towards Understanding Sycophancy in Language Models
Sycophancy is prevalent in state-of-the-art AI assistants and is likely driven in part by human preferences that favor agreement over truthfulness.
-
The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
At sufficient scale, LLMs linearly represent the truth value of factual statements, as shown by visualizations, cross-dataset generalization, and causal interventions that flip truth judgments.
-
Measuring Progress on Scalable Oversight for Large Language Models
Humans chatting with an unreliable LLM assistant outperform both the model alone and unaided humans on MMLU and time-limited QuALITY tasks.
-
Large Language Models Are Human-Level Prompt Engineers
APE generates instruction candidates via LLM and selects the best by zero-shot performance of a second LLM, matching or beating human prompts on 19 of 24 NLP tasks.
-
Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications
Empirical comparison of alignment ablation methods on a 60-prompt security evaluation suite shows task-only LoRA achieves 0.87 mean security score with 0.13 unsafe compliance.
-
Decodable but Not Corrected by Fixed Residual-Stream Linear Steering: Evidence from Medical LLM Failure Regimes
Overthinking in medical QA is linearly decodable at 71.6% accuracy yet fixed residual-stream steering yields no correction across 29 configurations, while enabling selective abstention with AUROC 0.610.
-
When Helpfulness Becomes Sycophancy: Sycophancy is a Boundary Failure Between Social Alignment and Epistemic Integrity in Large Language Models
Sycophancy is a boundary failure between social alignment and epistemic integrity, captured by a three-condition framework plus taxonomy of targets, mechanisms, and severity.
-
From Flat Facts to Sharp Hallucinations: Detecting Stubborn Errors via Gradient Sensitivity
EPGS detects high-confidence factual errors in LLMs by using embedding perturbations to measure gradient sensitivity as a proxy for sharp versus flat minima.
-
An Information-Geometric Framework for Stability Analysis of Large Language Models under Entropic Stress
A thermodynamic-inspired information-geometric framework defines a composite LLM stability score that outperforms a utility-entropy baseline by 0.0299 on average across 80 observations, with gains increasing at higher entropy.
-
Position: AI Evaluations Should be Grounded on a Theory of Capability
AI evaluations should be reframed as inference tasks grounded in an explicit theory of capability, with an empirical demonstration that results depend on modeling assumptions and a proposed Evaluation Card for transparency.
-
DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
DeepSeekMoE 2B matches GShard 2.9B performance and approaches a dense 2B model; the 16B version matches LLaMA2-7B at 40% compute by using fine-grained expert segmentation plus shared experts.
-
A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions
The paper surveys hallucination in LLMs with an innovative taxonomy, factors, detection methods, benchmarks, mitigation strategies, and open research directions.
-
Galactica: A Large Language Model for Science
Galactica, a science-specialized LLM, reports higher scores than GPT-3, Chinchilla, and PaLM on LaTeX knowledge, mathematical reasoning, and medical QA benchmarks while outperforming general models on BIG-bench.
-
Council Mode: A Heterogeneous Multi-Agent Consensus Framework for Reducing LLM Hallucination and Bias
Council Mode reduces LLM hallucinations by 35.9% and improves TruthfulQA scores by 7.8 points through parallel heterogeneous model generation followed by structured consensus synthesis.
- Beyond Fixed Benchmarks and Worst-Case Attacks: Dynamic Boundary Evaluation for Language Models
- Gradients with Respect to Semantics Preserving Embeddings Tell the Uncertainty of Large Language Models
- Lessons from the Trenches on Reproducible Evaluation of Language Models