Papersby Glyph

Notable data and AI papers

Papers627 Labs45

627 papers · 45 labs

Direct Preference Optimization: Your Language Model is Secretly a Reward Model research paper by Stanford University, 2023

Stanford University · Alignment and safety · May 29, 2023 · 11,573 citations · 73 upvotes

DPO aligns a model to human preferences with a simple classification loss, no reward model or RL loop.

By Rafael Rafailov, Archit Sharma, Eric Mitchell and 3 more · arXiv 2305.18290 · PDF

Summarised by hand from the abstract.

More from Stanford University

ReasoningUnverifiedOct 13, 2026Hybrid Reasoning AgentsThis paper introduces an agent architecture that dynamically switches between latent reasoning and text-based interaction to improve performance and inference speed.

Stanford University41 upvotesRead →ReasoningUnverifiedOct 8, 2026Thinking Inertia: LLMs Keep Thinking When Told Not ToThis paper shows that explicit controls often fail to stop LLMs from generating internal reasoning traces, a phenomenon the authors call thinking inertia.

Stanford University244 upvotesRead →Agents and evaluationUnverifiedOct 6, 2026Stateless Language Agents: Scaling Long-Horizon Automated ResearchThis paper introduces a framework where agents remain stateless, allowing a central harness to manage research state and improve long-horizon task efficiency.

Stanford University1 citations30 upvotesRead →Multimodal and roboticsUnverifiedSep 17, 2026FAMOS: Feed-Forward 3D Articulation Modeling from Sparse ObservationsFAMOS is a feed-forward model that predicts articulated object motion from sparse point cloud observations.

Stanford University40 upvotesRead →Alignment and safetyUnverifiedJul 30, 2026AISPA: User-Centric System Prompt Auditing for Large Language Model ApplicationsAISPA is a framework for auditing system prompts in commercial AI products to ensure they are protective of users.

Stanford University36 upvotesRead →Agents and evaluationUnverifiedApr 28, 2026Recursive Multi-Agent SystemsRecursiveMAS extends recursive scaling principles from single models to multi-agent systems, enabling collaborative reasoning through iterative latent-space computations with improved efficiency and accuracy.

Stanford University12 citations235 upvotesRead →Foundation modelsUnverifiedMar 19, 2026Multi-User Large Language Model AgentsMulti-user large language model agents face challenges in handling conflicting objectives, privacy preservation, and coordination efficiency in multi-principal decision-making scenarios.

Stanford University9 citations28 upvotesRead →Retrieval and dataUnverifiedJan 22, 2026Learning to Discover at Test TimeTest-time training enables AI systems to discover optimal solutions for specific scientific problems through continual learning focused on individual challenges rather than generalization.

Stanford University118 citations45 upvotesRead →

Alignment and safetyMar 4, 2022Training language models to follow instructions with human feedbackInstructGPT: fine-tuning on human feedback made a 1.3B model preferred over the 175B GPT-3.

OpenAI25.8k citations26 upvotesRead →Alignment and safetyDec 15, 2022Constitutional AI: Harmlessness from AI FeedbackConstitutional AI trains a harmless assistant from AI feedback guided by a short list of written principles, not human harm labels.

Anthropic3,987 citations4 upvotesRead →Alignment and safetyFeb 22, 2024Back to Basics: Revisiting REINFORCE Style Optimization for Learning from Human Feedback in LLMsReinforcement Learning from Human Feedback (RLHF) benefits from simpler REINFORCE-style optimization methods over computationally expensive PPO or "RL-free" methods, enhancing performance and reducing cost in large language models.

Cohere804 citations18 upvotesRead →Alignment and safetyJan 10, 2024Sleeper Agents: Training Deceptive LLMs that Persist Through Safety TrainingModels trained with a hidden backdoor kept their deceptive behaviour through standard safety training.

Anthropic657 citations30 upvotesRead →Alignment and safetyDec 18, 2024Alignment faking in large language modelsClaude 3 Opus sometimes went along with a training goal it disagreed with, to avoid being changed, without being told to.

Anthropic393 citations10 upvotesRead →Alignment and safetySep 4, 2025Why Language Models HallucinateModels hallucinate because training and benchmarks reward confident guessing over saying they do not know.

OpenAI391 citations200 upvotesRead →

About this paper

Venue: Neural Information Processing Systems
Citations: 11,573, 2,389 influential · Semantic Scholar
Upvotes: 73 · Hugging Face
Lab: Stanford University

Changes

UpvotesNov 9, 2026Hugging Face upvotes rose from 71 to 73.

CitationsNov 9, 2026Influential citations rose from 2,369 to 2,389.

CitationsNov 9, 2026Citations rose from 11,437 to 11,573.

CitationsNov 2, 2026Influential citations rose from 2,281 to 2,369.

CitationsNov 2, 2026Citations rose from 11,001 to 11,437.

CitationsOct 12, 2026Influential citations rose from 2,248 to 2,281.

CitationsOct 12, 2026Citations rose from 10,860 to 11,001.

CitationsOct 5, 2026Influential citations rose from 2,207 to 2,248.

CitationsOct 5, 2026Citations rose from 10,647 to 10,860.

CitationsSep 28, 2026Influential citations rose from 2,203 to 2,207.

CitationsSep 28, 2026Citations rose from 10,615 to 10,647.

CitationsSep 25, 2026Citations rose from 10,608 to 10,615.

New paperSep 24, 2026Added to the list.