{"list":"other","total":121,"rows":[{"id":"2005.14165","title":"Language Models are Few-Shot Learners","created":"2020-05-28","cited_by_count":32419},{"id":"2303.08774","title":"GPT-4 Technical Report","created":"2023-03-15","cited_by_count":18396},{"id":"2407.21783","title":"The Llama 3 Herd of Models","created":"2024-07-31","cited_by_count":14715},{"id":"1412.6572","title":"Explaining and Harnessing Adversarial Examples","created":"2014-12-20","cited_by_count":7364},{"id":"2001.08361","title":"Scaling Laws for Neural Language Models","created":"2020-01-23","cited_by_count":6586},{"id":"1312.6199","title":"Intriguing properties of neural networks","created":"2013-12-21","cited_by_count":5155},{"id":"2108.07258","title":"On the Opportunities and Risks of Foundation Models","created":"2021-08-16","cited_by_count":4033},{"id":"2204.05862","title":"Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback","created":"2022-04-12","cited_by_count":3720},{"id":"2307.15043","title":"Universal and Transferable Adversarial Attacks on Aligned Language Models","created":"2023-07-27","cited_by_count":3104},{"id":"2212.08073","title":"Constitutional AI: Harmlessness from AI Feedback","created":"2022-12-15","cited_by_count":3050},{"id":"2109.07958","title":"TruthfulQA: Measuring How Models Mimic Human Falsehoods","created":"2021-09-08","cited_by_count":2975},{"id":"2206.07682","title":"Emergent Abilities of Large Language Models","created":"2022-06-15","cited_by_count":2512},{"id":"2206.04615","title":"Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models","created":"2022-06-09","cited_by_count":1998},{"id":"2112.09332","title":"WebGPT: Browser-assisted question-answering with human feedback","created":"2021-12-17","cited_by_count":1745},{"id":"2310.01405","title":"Representation Engineering: A Top-Down Approach to AI Transparency","created":"2023-10-02","cited_by_count":1336},{"id":"2310.03693","title":"Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!","created":"2023-10-05","cited_by_count":1154},{"id":"2408.06292","title":"The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery","created":"2024-08-12","cited_by_count":1071},{"id":"2210.07229","title":"Mass-Editing Memory in a Transformer","created":"2022-10-13","cited_by_count":1001},{"id":"2212.09251","title":"Discovering Language Model Behaviors with Model-Written Evaluations","created":"2022-12-19","cited_by_count":836},{"id":"2310.11324","title":"Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formatting","created":"2023-10-17","cited_by_count":817},{"id":"2202.07646","title":"Quantifying Memorization Across Neural Language Models","created":"2022-02-15","cited_by_count":763},{"id":"2307.13702","title":"Measuring Faithfulness in Chain-of-Thought Reasoning","created":"2023-07-17","cited_by_count":703},{"id":"2404.13076","title":"LLM Evaluators Recognize and Favor Their Own Generations","created":"2024-04-15","cited_by_count":686},{"id":"2010.09670","title":"RobustBench: a standardized adversarial robustness benchmark","created":"2020-10-19","cited_by_count":661},{"id":"2308.08747","title":"An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning","created":"2023-08-17","cited_by_count":639},{"id":"2401.05566","title":"Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training","created":"2024-01-10","cited_by_count":529},{"id":"1908.09203","title":"Release Strategies and the Social Impacts of Language Models","created":"2019-08-24","cited_by_count":513},{"id":"2403.03218","title":"The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning","created":"2024-03-05","cited_by_count":509},{"id":"2312.09390","title":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision","created":"2023-12-14","cited_by_count":467},{"id":"2104.14337","title":"Dynabench: Rethinking Benchmarking in NLP","created":"2021-04-07","cited_by_count":459},{"id":"2210.01504","title":"Knowledge Unlearning for Mitigating Privacy Risks in Language Models","created":"2022-10-04","cited_by_count":437},{"id":"2404.14082","title":"Mechanistic Interpretability for AI Safety -- A Review","created":"2024-04-22","cited_by_count":422},{"id":"1805.00899","title":"AI safety via debate","created":"2018-05-02","cited_by_count":400},{"id":"2309.07875","title":"Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions","created":"2023-09-14","cited_by_count":387},{"id":"2310.18018","title":"NLP Evaluation in trouble: On the Need to Measure LLM Data Contamination for each Benchmark","created":"2023-10-27","cited_by_count":348},{"id":"2305.13300","title":"Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts","created":"2023-05-22","cited_by_count":341},{"id":"2110.01963","title":"Multimodal datasets: misogyny, pornography, and malignant stereotypes","created":"2021-10-05","cited_by_count":327},{"id":"2310.02446","title":"Low-Resource Languages Jailbreak GPT-4","created":"2023-10-03","cited_by_count":309},{"id":"1801.09344","title":"Certified Defenses against Adversarial Examples","created":"2018-01-29","cited_by_count":306},{"id":"2310.02949","title":"Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models","created":"2023-10-04","cited_by_count":297},{"id":"2109.13916","title":"Unsolved Problems in ML Safety","created":"2021-09-28","cited_by_count":292},{"id":"2301.13379","title":"Faithful Chain-of-Thought Reasoning","created":"2023-01-31","cited_by_count":288},{"id":"2402.06782","title":"Debating with More Persuasive LLMs Leads to More Truthful Answers","created":"2024-02-09","cited_by_count":266},{"id":"2209.00626","title":"The Alignment Problem from a Deep Learning Perspective","created":"2022-08-30","cited_by_count":258},{"id":"2312.06942","title":"AI Control: Improving Safety Despite Intentional Subversion","created":"2023-12-12","cited_by_count":229},{"id":"2306.04528","title":"PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Prompts","created":"2023-06-07","cited_by_count":226},{"id":"2211.03540","title":"Measuring Progress on Scalable Oversight for Large Language Models","created":"2022-11-04","cited_by_count":225},{"id":"2310.05492","title":"How Abilities in Large Language Models are Affected by Supervised Fine-tuning Data Composition","created":"2023-10-09","cited_by_count":223},{"id":"2012.15761","title":"Learning from the Worst: Dynamically Generated Datasets to Improve Online Hate Detection","created":"2020-12-31","cited_by_count":221},{"id":"2309.07045","title":"SafetyBench: Evaluating the Safety of Large Language Models","created":"2023-09-13","cited_by_count":199},{"id":"2404.09932","title":"Foundational Challenges in Assuring Alignment and Safety of Large Language Models","created":"2024-04-15","cited_by_count":199},{"id":"2310.20624","title":"LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B","created":"2023-10-31","cited_by_count":195},{"id":"2310.10844","title":"Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks","created":"2023-10-16","cited_by_count":194},{"id":"2311.01964","title":"Don't Make Your LLM an Evaluation Benchmark Cheater","created":"2023-11-03","cited_by_count":194},{"id":"2406.10162","title":"Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models","created":"2024-06-14","cited_by_count":178},{"id":"2305.15324","title":"Model evaluation for extreme risks","created":"2023-05-24","cited_by_count":177},{"id":"2309.00236","title":"Image Hijacks: Adversarial Images can Control Generative Models at Runtime","created":"2023-09-01","cited_by_count":172},{"id":"1911.03842","title":"Queens are Powerful too: Mitigating Gender Bias in Dialogue Generation","created":"2019-11-10","cited_by_count":152},{"id":"2406.04244","title":"Benchmark Data Contamination of Large Language Models: A Survey","created":"2024-06-06","cited_by_count":147},{"id":"2309.00667","title":"Taken out of context: On measuring situational awareness in LLMs","created":"2023-09-01","cited_by_count":143},{"id":"2305.10160","title":"Stop Uploading Test Data in Plain Text: Practical Strategies for Mitigating Data Contamination by Evaluation Benchmarks","created":"2023-05-17","cited_by_count":142},{"id":"2312.04724","title":"Purple Llama CyberSecEval: A Secure Coding Benchmark for Language Models","created":"2023-12-07","cited_by_count":138},{"id":"2408.00761","title":"Tamper-Resistant Safeguards for Open-Weight LLMs","created":"2024-08-01","cited_by_count":138},{"id":"2403.09792","title":"Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models","created":"2024-03-14","cited_by_count":136},{"id":"2304.05332","title":"Emergent autonomous scientific research capabilities of large language models","created":"2023-04-11","cited_by_count":132},{"id":"2211.09110","title":"Holistic Evaluation of Language Models","created":"2022-11-16","cited_by_count":131},{"id":"2404.08144","title":"LLM Agents can Autonomously Exploit One-day Vulnerabilities","created":"2024-04-11","cited_by_count":128},{"id":"2210.01790","title":"Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals","created":"2022-10-04","cited_by_count":126},{"id":"2108.01624","title":"Large-Scale Differentially Private BERT","created":"2021-08-03","cited_by_count":118},{"id":"2305.14710","title":"Instructions as Backdoors: Backdoor Vulnerabilities of Instruction Tuning for Large Language Models","created":"2023-05-24","cited_by_count":110},{"id":"2312.09085","title":"The Earth is Flat because...: Investigating LLMs' Belief towards Misinformation via Persuasive Conversation","created":"2023-12-14","cited_by_count":109},{"id":"2403.14380","title":"On the Conversational Persuasiveness of Large Language Models: A Randomized Controlled Trial","created":"2024-03-21","cited_by_count":109},{"id":"2404.18824","title":"Benchmarking Benchmark Leakage in Large Language Models","created":"2024-04-29","cited_by_count":106},{"id":"2307.03718","title":"Frontier AI Regulation: Managing Emerging Risks to Public Safety","created":"2023-07-06","cited_by_count":98},{"id":"2312.11671","title":"Evaluating Language-Model Agents on Realistic Autonomous Tasks","created":"2023-12-18","cited_by_count":96},{"id":"2305.13160","title":"Can ChatGPT Defend its Belief in Truth? Evaluating LLM Reasoning via Debate","created":"2023-05-22","cited_by_count":94},{"id":"2403.15796","title":"Understanding Emergent Abilities of Language Models from the Loss Perspective","created":"2024-03-23","cited_by_count":94},{"id":"2306.05949","title":"Evaluating the Social Impact of Generative AI Systems in Systems and Society","created":"2023-06-09","cited_by_count":92},{"id":"2403.17336","title":"Don't Listen To Me: Understanding and Exploring Jailbreak Prompts of Large Language Models","created":"2024-03-26","cited_by_count":86},{"id":"2304.12298","title":"BadGPT: Exploring Security Vulnerabilities of ChatGPT via Backdoor Attacks to InstructGPT","created":"2023-02-21","cited_by_count":76},{"id":"2203.04592","title":"Mapping global dynamics of benchmark creation and saturation in artificial intelligence","created":"2022-03-09","cited_by_count":71},{"id":"1712.02950","title":"CycleGAN, a Master of Steganography","created":"2017-12-08","cited_by_count":65},{"id":"2310.18512","title":"Preventing Language Models From Hiding Their Reasoning","created":"2023-10-27","cited_by_count":65},{"id":"2012.15349","title":"DynaSent: A Dynamic Benchmark for Sentiment Analysis","created":"2020-12-30","cited_by_count":63},{"id":"2407.21792","title":"Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?","created":"2024-07-31","cited_by_count":61},{"id":"2403.02817","title":"Here Comes The AI Worm: Unleashing Zero-click Worms that Target GenAI-Powered Applications","created":"2024-03-05","cited_by_count":59},{"id":"2402.04247","title":"Risks of AI Scientists: Prioritizing Safeguarding Over Autonomy","created":"2024-02-06","cited_by_count":49},{"id":"2403.10462","title":"Safety Cases: How to Justify the Safety of Advanced AI Systems","created":"2024-03-15","cited_by_count":48},{"id":"2403.04893","title":"A Safe Harbor for AI Evaluation and Red Teaming","created":"2024-03-07","cited_by_count":47},{"id":"2406.07882","title":"Designing a Dashboard for Transparency and Control of Conversational AI","created":"2024-06-12","cited_by_count":46},{"id":"2311.15766","title":"Knowledge Unlearning for LLMs: Tasks, Methods, and Challenges","created":"2023-11-27","cited_by_count":38},{"id":"2405.01576","title":"Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistant","created":"2024-04-25","cited_by_count":38},{"id":"2406.17864","title":"AI Risk Categorization Decoded (AIR 2024): From Government Regulations to Corporate Policies","created":"2024-06-25","cited_by_count":37},{"id":"2311.00117","title":"BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B","created":"2023-10-31","cited_by_count":36},{"id":"2306.03809","title":"Can large language models democratize access to dual-use biotechnology?","created":"2023-06-06","cited_by_count":35},{"id":"2402.16382","title":"Immunization against harmful fine-tuning attacks","created":"2024-02-26","cited_by_count":35},{"id":"2311.09227","title":"Open-Sourcing Highly Capable Foundation Models: An evaluation of risks, benefits, and alternative methods for pursuing open-source objectives","created":"2023-09-29","cited_by_count":31},{"id":"2305.16519","title":"The Dangers of trusting Stochastic Parrots: Faithfulness and Trust in Open-domain Conversational Question Answering","created":"2023-05-25","cited_by_count":26},{"id":"2406.13236","title":"Data Contamination Can Cross Language Barriers","created":"2024-06-19","cited_by_count":26},{"id":"2407.12856","title":"AI-AI Bias: large language models favor communications generated by large language models","created":"2024-07-09","cited_by_count":26}],"next_cursor":"100"}