Flips

Follow this topic Ask about this topic

Momentum

7 papers in the last four weeks, up 75% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 48

All topics

How Divergence Becomes Decision Flips in Compressed Language Models
Sep 30, 2026Beatriz Almeida FelicioKullback-Leibler Divergence Large Language Model Decoding
Instituto de Informática (INF) Universidade Federal de Goiás (UFG)
SignFLIP: A Unified Model for Sign Language Translation and Generation via Stage-wise Alignment at Scale
Sep 28, 2026Zhaoyi An, Sihan Tan, Youngbae Hwang +2Gloss-Free Sign Language Translation Sign Language Translation
Institute of Science Tokyo · Chungbuk National University
JevOut: Natural Context Can Flip Decision Models
Sep 24, 2026Zixiang XuDecisions Contextual Reasoning
University of Southern California
FlipToSee: A Probabilistic Stable Placement Prior for Active Visual Exploration via Regrasping
Sep 17, 2026Chang Shu, Sushil Samuel Dinesh, Shinkyu ParkViewpoint-Dependent Active Perception Visual Place Recognition
Electrical and Computer Engineering, King Abdullah University of Science and Technology (KAUST), Thuwal, 23955-6900, Saudi Arabia
Overflip: Repetition-Induced Label Flips in Guardrail Models
Sep 14, 2026Xu He, Chih-Hsuan Lin, Hung-Mao Chen +3Flips Guardrail
1George Mason University · 2Virginia Tech · 3Missouri University of Science and Technology +1
Style Over Substance: Content-Invariant Wrappers Flip LLM Safety-Judge Verdicts
Sep 8, 2026Yongxi Zhou, Wenbo Ye, Yuanzhe Liu +2Large Language Model Safety Jailbreak Success Rates
Northeastern University, Massachusetts, USA · University of Southern California, California, USA · Georgia Institute of Technology, Georgia, USA
Quantization Amplifies Determinism, Not Bias: Scale-Dependent Behavioral Effects of Serving-Time Weight Compression
Sep 7, 2026Dachi KurtskhaliaLarge Language Model Quantization Quantized
Impel
Flip, Don't Shuffle: Watermarking LLMs at the Speed of Inference
Sep 3, 2026Simone Ceppi, Ignacio SanchezLarge Language Model Watermarks Watermarking
European Commission Joint Research Centre Ispra, Italy
Measuring Memory and Generalization as Separable Geometric Channels: The Topo^2 Framework
Aug 31, 2026Zhanbo Zhang, Ming Liu, Qing WangMemorization Topology
Anhui iFlytek Yinglian Technology Co., Ltd.
Build it, Break it, Repeat: Benchmarking and improving LLM-manipulated disinformation detection in social media posts
Aug 10, 2026Kevin Thomas, Milosz Kasprzyk, Reuel C Igbokwe Onuigbo +5Misinformation Detection Attacker Large Language Model
University of Sheffield, Department of Computer Science, Regent Court, 211 Portobello Street, Sheffield, S1 4DP, United Kingdom
Targeted Label-Flipping and Oversampling Attacks on Federated Conditional GANs
Aug 10, 2026Panav Shah, Avishek GhoshFederated Learning Black-Box Adversarial Attacks
Indian Institute of Technology, Bombay Mumbai, India
Which Decisions Low-Bit Quantization Breaks, and How to Predict Them
Aug 6, 2026Zekun Wu, Swati Dhiman, Adriano KoshiyamaQuantized Shrink
Holistic AI · UCL Centre for Artificial Intelligence, University College London
Test, then Route: How Language Models Execute In-Context Conditional Rules Across Models and Languages
Aug 4, 2026Luxshan Thavarasa, Sivasuthan SukumarFlips Switch
Independent Researcher, Colombo, Sri Lanka · Department of Computer Science and Engineering, University of Moratuwa, Sri Lanka
StanceFlip: A Comprehensive Multi-Dimensional Benchmark for Multimodal Conversational Stance Flipping Forecasting
Jul 27, 2026Heyan Chai, Xin Li, Wenjie Wang +5Multimodal Sentiment Analysis Flips
College of Computer Science and Software Engineering, Shenzhen University Shenzhen, China · Harbin Institute of Technology Shenzhen, China · Shenzhen University Shenzhen, China +1
Constraint-Aware Counterfactual Editing for Aspect-Based Sentiment Analysis
Jul 15, 2026S M Rafiuddin, Vamsi Krishna Pavuluri, Atriya SenCounterfactual Generation Sentiment
Department of Computer Science, Oklahoma State University, Stillwater, Oklahoma, USA
Learning to Throw Objects Safely in Multi-Obstacle Environments
Jul 7, 2026Mohammadreza Kasaei, Klemen Voncina, Hamidreza KasaeiRobotic Manipulation Obstacle Avoidance
School of Informatics, University of Edinburgh, UK · Department of Artificial Intelligence, University of Groningen, The Netherlands
A Coin Flip Per Token: Bernoulli Sparse Steering of Large Language Models
Jul 6, 2026Nima Eshraghi, Lovedeep Gondara, Yuqing Huang +5Linear Activation Steering Steering
1The Vanguard Group, Inc. · University of Toronto
NormWorlds-CF: Solver-Verified Counterfactual Normative Reasoning with Metamorphic-Relation GRPO
Jul 4, 2026Xinqi ZhangFlips
Tsinghua University
Boundary-Aware Quantization: Finite-Scale Decision Geometry of Neural Classifiers
Jul 1, 2026O. M. KiselevQuantized Decision Boundaries
Innopolis University, Innopolis, Republic of Tatarstan, Russian Federation
Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models
Jun 24, 2026Akshay Paruchuri, Sanmi Koyejo, Ehsan AdeliMultimodal Large Language Models Flips
Stanford University
Necessary but Not Sufficient: Temperature Control and Reproducibility in LLM-as-Judge Safety Evaluations
Jun 24, 2026Hiroki TambaReproducibility Disagreement
Tamba Research Academy
Transferable Attack against Face Swapping in an Extended Space
Jun 24, 2026Mingzhi Lyu, Yi Huang, Jun Xie +3Deepfake Detection Flips
Nanyang Technological University Singapore
Heterogeneous LLM Debate Under Adversarial Peers: Honest Gains, Replacement Costs, and Resilience
Jun 18, 2026Prashanti Nilayam, Kiran Kumar Ramanna, Prashil Tumbade +1Adversarial Robustness Honesty
ServiceNow
Semantic Flip: Synthetic OOD Generation for Robust Refusal in Embodied Question Answering and Spatial Localization
Jun 15, 2026Dongbin Na, Chanwoo Kim, Giyun Choi +1Refusals Flips
RGA Inc.
Who Flips? Self- and Cross-Model Counterarguments Reveal Answer Instability in LLMs
Jun 14, 2026Nafiseh Nikeghbal, Amir Hossein Kargaran, Shaghayegh Kolli +1Flips Answer
1Technical University of Munich · Munich Center for Machine Learning · 1Technical University of Munich 2 Munich Center for Machine Learning +2
CoCoGEC: Counterfactual Generation for Robust Grammatical Error Correction
Jun 13, 2026Qianyu Wang, Xiaoman Wang, Yuanyuan Liang +2Grammatical Error Correction Counterfactual Generation
East China Normal University
Model Poisoning Against Federated Model Adaptation with Chain of Bit-Flips
Jun 8, 2026Bastien Vuillod, Kevin Hector, Pierre-Alain Moellic +2Federated Learning Poisoning
CEA-Leti, Mines Saint-Etienne, Equipe Commune SAS, F-13541 Gardanne, France · Mines Saint-Etienne, CEA-Leti, Centre CMP, Equipe commune SAS, F-13541 Gardanne, France
Vision-Language Asymmetry in Bistable Image Captioning
Jun 6, 2026Arohan AgateImage Difference Captioning Text Dominance
University of Washington.
VisualFLIP: Do Predictions Depend on Task-Critical Visual Evidence in Multimodal Reasoning?
Jun 5, 2026Didi Zhu, Changrui Chen, Stefanos Zafeiriou +1Multimodal Reasoning Benchmarks Multimodal Reasoning
Imperial College London
Decomposing Factual Sycophancy in Language Models: How Size and Instruction Tuning Shape Robustness
Jun 4, 2026Victor De Marez, Luna De Bruyne, Walter DaelemansSycophancy Instruction-Tuned Models
Centre for Computational Linguistics, Psycholinguistics and Sociolinguistics University of Antwerp Antwerp, Belgium
TAM: Torque Adaptation Module for Robust Motion Transfer in Manipulation
Jun 4, 2026Dongwon Son, Florian Shkurti, Jason Lee +3Torque Robotic Manipulation Policies
1KAIST · *Work done during an internship at the Allen Institute for AI. · 2Allen Institute for AI +2
FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences
Jun 2, 2026Gurvan Richardeau, Gohar Dashyan, Erwan Le Merrer +1Large Language Model Safety Large Language Model Baselines
PEReN (Paris, France) · Université de Rennes, Inria, CNRS/IRISA (Rennes, France) · LAAS, CNRS, (Toulouse, France)
FlipItRight: Stable Pose-Targeted Throw-Flip Across Diverse Objects
Jun 1, 2026Axel Dawne, Shinkyu ParkRobotic Manipulator Category-Level Object Pose Estimation
King Abdullah University of Science and Technology (KAUST), Saudi Arabia
The Shape of Wisdom: Decision Trajectories in Language Models
May 31, 2026Shailesh RanaFlips Multilayer Perceptrons
Independent Researcher
TriSearch: Learning to Optimize Triangulations via Bistellar Flips
May 28, 2026Yiran Wang, Guido MontúfarTriangulations Polytopes
UCLA · MPI MiS
MarginGate: Sparse Margin-Triggered Verification for Batch-Invariant LLM Inference
May 28, 2026Kexin Chu, Yang Zhou, Wei ZhangLLM Inference Optimization Decoding
University of Connecticut · UC Davis
When Context Flips, Safety Breaks: Diagnosing Brittle Safety in Aligned Language Models
May 27, 2026Dasol Choi, Alex KwonLarge Language Model Safety Safety Benchmarks
AIM Intelligence · Independent Researcher
Constraint-Anchored Attribution: Feasibility-Certified Counterfactuals and Bonferroni-PAC Sufficient Subsets for Neural CO Policies
May 24, 2026Sohaib LafifiNeural Combinatorial Optimization Credit Assignment
Univ. Artois, UR 3926, Laboratoire de Génie Informatique et d’Automatique de l’Artois (LGI2A) Béthune, France
Pitfalls of Unlabeled Disagreement-Based Drift Detection in Streaming Tree Ensembles
May 12, 2026Lara Sá Neves, Afonso Lourenço, Lizy K. John +1Tree Ensembles Unlabeled Data
GECAD, ISEP, Polytechnic of Porto, Portugal · The University of Texas at Austin, USA
Simulating Students or Sycophantic Problem Solving? On Misconception Faithfulness of LLM Simulators
May 12, 2026Heejin Do, Shashank Sonkar, Mrinmaya SachanUser Simulation Student Misconceptions
ETH Zürich, ETH AI Center · University of Central Florida · ETH Zürich
Intervention-Based Time Series Causal Discovery via Simulator-Generated Interventional Distributions
May 11, 2026Tsuyoshi OkitaCausal Discovery Methods Causal
Kyushu Institute of Technology
Efficient Preference Poisoning Attack on Offline RLHF
May 4, 2026Chenye Yang, Weiyu Xu, Lifeng LaiBacktranslation Augmented Direct Preference Optimization Reinforcement Learning From Human Feedback
Department of Electrical and Computer Engineering, University of California, Davis, Davis, CA, USA · Department of Electrical and Computer Engineering, University of Iowa, Iowa City, IA, USA.
Compared to What? Baselines and Metrics for Counterfactual Prompting
May 1, 2026Zihao Yang, Mosh Levy, Yoav Goldberg +1Large Language Model Bias Causal Intervention
Northeastern University · Bar-Ilan University · Allen Institute for AI
Towards High Fidelity Face Swapping: A Comprehensive Survey and New Benchmark
Apr 27, 2026Qi Li, Weining Wang, Shuangjun Du +5Deepfake Detection Generative Adversarial Network
New Laboratory of Pattern Recognition (NLPR), State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), Institute of Automation, Chinese Academy of Sciences, Beijing, 100190, China · School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100049, China · Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, 100190, China +3
The Coin Flip Judge? Reliability and Bias in LLM-as-a-Judge Evaluation
Apr 23, 2026Abel YagubyanLlm-As-A-Judge Large Language Model Judges
Independent Researcher
Prompt Optimization Is a Coin Flip: Diagnosing When It Helps in Compound AI Systems
Apr 16, 2026Xing Zhang, Guanghui Wang, Yanwei Cui +4Prompt Optimization Instruction-Tuned Models
AWS Generative AI Innovation Center · HSBC Holdings Plc., HSBC Technology Center, China.
From Arithmetic to Logic: The Resilience of Logic and Lookup-Based Neural Networks Under Parameter Bit-Flips
Mar 24, 2026Alan T. L. Bacellar, Sathvik Chemudupati, Shashank Nag +4Differentiable Logic Gate Networks Neural Network
The University of Texas at Austin · Federal University of Rio de Janeiro · Instituto de Telecomunicações, Porto, Portugal
Approaching the Harm of Gradient Attacks While Only Flipping Labels
Feb 28, 2025Abdessamad El-Kabid, El-Mahdi El-MhamdiFederated Learning Flips
École Polytechnique

Sep 30, 2026, cs.CLJ/K move · Enter open · S save

How Divergence Becomes Decision Flips in Compressed Language Models

Beatriz Almeida Felicio

Instituto de Informática (INF) Universidade Federal de Goiás (UFG)

Kullback-Leibler Divergence Large Language Model Decoding Divergence Flips Report

Open paper AskPDF arXiv

Compression reports summarize how far a compressed language model moved from the dense one, usually by a KL divergence; a deployment that relies on the dense model's outputs needs to know how many of its decisions changed. We show that total variation, not KL, answers this directly. Across 802 compressed and perturbed copies of 19 open models on five corpora and nine mechanically unrelated perturbation families, the rate at which the arg-max token changes (the \emph{flip rate}) tracks total variation at a ratio with median

1.05

, with no fitted constant. KL converts into flips only through its square root and a factor that varies fourfold across models and corpora, because KL averages over tokens before the root is taken; first-order statistics averaged per token, such as Hellinger distance, avoid this, but reports rarely give them. As a result, of two compressors reported on different models and corpora whose flip rates differ by at least

10%

, KL assigns the smaller divergence to the one that changes more decisions in

11%

of cases, total variation in

1%

. Two pre-registered tests mark the limits: on a held-out code corpus the ratio held for all eight models while three predictions about KL each failed for half of them or more, and on three new models with real kernels it stayed in its band for 37 of 38 checkpoints but fell below one on code for two models. In vLLM speculative decoding, total variation measured under teacher forcing predicts greedy draft acceptance with a mean relative error of

1.1

2.4%

, without the task-specific calibration that KL needs.

Momentum

Related topics

Latest papers 48