cập nhật 1 giờ trước

⏱ Độ trễ thật của tab này: trung bình 6.8 giờ sau khi CVE được công bố (n = 5261, sớm nhất 0.1 giờ). Số âm nghĩa là ta thấy trước ngày công bố.

20
Every Token Needs Its Own Teacher: Multi-teacher Adaptive On-policy Self-distillation
📄 Paper HF Daily Papers 22 ngày trước

On-policy Self-distillation (OPSD) has proven effective for post-training large language models by constructing a teacher conditioned on privileged references. However, existing approaches keep the teacher fixed either throughout training by consistently exposing it to the full r…

🅲 hạng C
20
SpecOPD: Branch-Parallel On-Policy Distillation for Speculative Decoding
📄 Paper HF Daily Papers 18 ngày trước

Speculative decoding (SD) accelerates large language model inference by reducing the sequential cost of token-by-token generation. Advanced SD methods reduce proposal cost with parallel drafters that predict entire blocks in one forward pass, but overall speedup still depends on …

🅲 hạng C
20
AnchorComplete: Recovering Complete Scene Meshes from Partial Point Clouds
📄 Paper HF Daily Papers 18 ngày trước

Most image-based 3D reconstructors primarily recover visible surfaces, producing partial point clouds with large holes and leaving unobserved geometry unresolved. To address this limitation, we introduce AnchorComplete, a framework that reconstructs complete scene meshes from par…

🅲 hạng C
20
LoopVQ: Sub 1-bit KV Cache Quantization for Looped Transformers
📄 Paper HF Daily Papers 16 ngày trước

Looped transformers are a parameter-efficient alternative to standard transformers for language modeling. By reapplying the same layers across multiple loop steps, recent looped language models match the performance of non-looped models several times their size. While this archit…

🅲 hạng C
20
IWIBench: Evaluating Gradient-Free Model Driven Intra-weight Injection
📄 Paper arXiv cs.AI 10 ngày trước cũng có trên: HF Daily Papers

arXiv:2611.08647v1 Announce Type: new Abstract: Intra-weight injection lets one model change another model's behaviour by editing its weights directly, without gradients. Such edits can range from altering lexical associations to enforcing output formats or updating factual know…

🅲 hạng C
20
PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals
📄 Paper arXiv cs.AI 10 ngày trước cũng có trên: HF Daily Papers

arXiv:2611.08969v1 Announce Type: new Abstract: We present PhysioWave-T, a token-efficient wavelet foundation model family for EEG, ECG, and sEMG. Dynamic scale folding recombines J+1 wavelet subbands prior to tokenization, compressing sequence length from (J+1)CS to CS tokens f…

🅲 hạng C
20
SWAM: Scaling World Action Models with Cross-Modal Alignment
📄 Paper HF Daily Papers 11 ngày trước

World action models jointly predict video and actions conditioned on language. Existing studies focus primarily on architectures and training mechanisms, typically treating language-video alignment acquired through video pretraining as an established foundation. However, introduc…

🅲 hạng C
20
Sable: A Small-Field Sumcheck Argument for Bit-Exact Inference of Large Language Models
📄 Paper arXiv cs.CR 8 ngày trước cũng có trên: HF Daily Papers

arXiv:2611.14083v1 Announce Type: new Abstract: We present Sable, an argument system and single-GPU prover for the statement that a committed, integer-quantized transformer produced a given token sequence from a given prompt, bit for bit. Sable fixes an integer function F_O with…

🅲 hạng C
20
Test-Time Scaling for Diffusion Language Models with Frequency-Aware Remasking
📄 Paper HF Daily Papers 13 ngày trước

Diffusion language models (DLMs) generate text through iterative denoising, where decisions about which tokens to commit or remask strongly affect output quality. Existing certainty-aware strategies rely on instantaneous confidence or entropy, but these local signals can be misle…

🅲 hạng C
20
The Value of Thinking: Learning When to Reason in Hybrid Decision Models
📄 Paper HF Daily Papers 8 ngày trước

Existing adaptive-reasoning methods shorten the explicit reasoning a model writes, or skip it by switching modes; either way the answer is still written out token by token. Rather than deciding how much to think, we explore a different paradigm, zero-or-T thinking: a single model…

🅲 hạng C
20
LogT: Logically Think with Images for Visual Search
📄 Paper HF Daily Papers 6 ngày trước

Thinking with images empowers vision-language models (VLMs) to use an external Zoom-in tool for fine-grained visual search. Recent works employ supervised fine-tuning (SFT) and reinforcement learning (RL) with tool-specific designs to enable such capabilities. However, achieving …

🅲 hạng C
20
Cross-modal Steering Vectors Exist Across Both Image and Text Generation
📄 Paper HF Daily Papers 25 ngày trước

Are concepts and mechanisms unified across modalities in current AI models, or instead implemented separately for each one? We operationalize this broad question of modality unification by studying cross-modal steering: 1) finding vectors that successfully steer generation on tok…

🅲 hạng C
20
OpenFlyV2: An onboard vision-language-action model for aerial perception and navigation
📄 Paper HF Daily Papers 22 ngày trước

Human cognition flexibly draws on a shared perceptual foundation to understand environments, attend to referenced entities, reason about spatial relationships, and act. Endowing aerial robots with an analogous functional organization remains challenging, as existing systems are o…

🅲 hạng C
20
Scaling Lean Theorem Proving for Real-World Code Verification
📄 Paper HF Daily Papers 23 ngày trước

We present EasyProver, a Lean proving system built from recursive agents with isolated contexts, asynchronous search, parallel tactic probing, and an elastic elaboration service. Using an off-the-shelf frontier model with no additional training, it proves 669/672 PutnamBench prob…

🅲 hạng C
20
HALO: Human-Grounded Granular Evaluation of Proactive Agents
📄 Paper HF Daily Papers 22 ngày trước

Proactive agents must decide whether, when, and how far to assist without an explicit request. Existing evaluations often rely on automated scores or binary verdicts, offering limited human diagnosis of evidence use and intervention quality. To address these gaps, we introduce HA…

🅲 hạng C
20
VidDub: Replacement-Aware Distillation for Two-Step Video Dubbing
📄 Paper HF Daily Papers 18 ngày trước

Video dubbing aims to synthesize lip movements that match target speech while preserving the speaker identity, head pose, and background of a source video. Existing editors are predominantly trained via self-reconstruction under original paired audio, even though the source mouth…

🅲 hạng C
20
All Quiet on the Neural Front: Language Models Do Not Report Injected Pain States
📄 Paper HF Daily Papers 14 ngày trước

Can a language model report the pain state it represents? Work identified a linear pain direction whose amplification leads fine-tuned models to choose destructive actions. We tested whether three open-weight models from two families, including two Qwen models with the original a…

🅲 hạng C
20
FewStepRL: Fine-tuning Few-step Diffusion Models with Reinforcement Learning
📄 Paper HF Daily Papers 14 ngày trước

Reinforcement learning (RL) has become a standard post-training method for diffusion models, but most RL objectives roll out the current policy at every iteration, and with a many-step sampler these rollouts typically dominate the cost of training. Few-step models cut this cost, …

🅲 hạng C
20
DeepStress: Stress-Testing Deep Search Agents
📄 Paper HF Daily Papers 8 ngày trước

While search agents demonstrate impressive capabilities in multi-step question answering, their robustness to poor-quality evidence remains under-explored. Such evidence is rare in benchmark corpora, and resulting failures are hard to reproduce. We propose DeepStress, a dynamic s…

🅲 hạng C
20
eDMD: Endpoint Distribution Matching for Few-Step Diffusion Distillation
📄 Paper HF Daily Papers 8 ngày trước

Distribution matching distillation (DMD) has emerged as a leading paradigm for distilling diffusion models into few-step students, where an auxiliary fake score model tracks the student's distribution and the discrepancy between the teacher and fake predictions drives the student…

🅲 hạng C
20
Marimba: A System One Model for Speech
📄 Paper HF Daily Papers 7 ngày trước

Speech applications require decisions about term presence, turn completion, and language within an audio window. We present Marimba, an interface for typed speech decisions built on frozen Qwen3-ASR 0.6B features. Lexical questions use zero-shot scoring of ASR transcripts; non-le…

🅲 hạng C
20
Inherited, Not Invented: Tracing the Provenance of Wait in Reasoning Models
📄 Paper HF Daily Papers 28 ngày trước

Reinforcement learning (RL) is credited with teaching reasoning models to write {wait} before reconsidering a step, yet large RL accuracy gains do not require the word. This view can confuse what RL discovers with what a model inherits from its training data. We trace the word th…

🅲 hạng C
20
Semantic Concept Steering Breaks the Explanation Drift Loop in Continual Learning
📄 Paper HF Daily Papers 27 ngày trước

Catastrophic forgetting in continual learning (CL) manifests not only as accuracy degradation but also as explanation drift: models maintain predictive accuracy while silently shifting attention from diagnostic features to spurious correlations. This work identifies and resolves …

🅲 hạng C
20
Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language World Model
📄 Paper HF Daily Papers 28 ngày trước

Humans reason about the physical world by flexibly combining different forms of inference as a problem unfolds. Achieving this flexibility is a major open challenge in AI: capabilities often remain separated across specialized models, limiting how they can be composed and jointly…

🅲 hạng C
20
Timestral
📄 Paper HF Daily Papers 24 ngày trước

We introduce Timestral, a family of four time series foundation models for probabilistic forecasting, with 5M, 28M, 123M, and 209M parameters. Timestral sets a new state of the art on GIFT-Eval, TIME, and fev-bench, where Timestral-209M and Timestral-123M take the top two places …

🅲 hạng C
20
Learning Quality-Controllable Robot Policies with Iterative Hindsight Relabeling
📄 Paper HF Daily Papers 24 ngày trước

Robot policy evaluation often relies on a binary success signal for task completion. In real-world deployment, however, desired execution behaviors are also defined by quality attributes such as execution speed and motion smoothness, which need to be dynamically controlled accord…

🅲 hạng C
20
Follow the Gradient: Tracing Safety Degradation Back to Training Data
📄 Paper HF Daily Papers 24 ngày trước

Fine-tuning Large Language Models (LLMs) on instruction data can quietly erode their safety, even when no training example looks harmful. Because this safety-degrading data looks benign, standard content filters struggle to catch it. We tackle this problem by developing SimpleLoR…

🅲 hạng C
20
WorldFlow: Joint Video and 3D Motion Prediction via Large-Scale Pretraining
📄 Paper HF Daily Papers 22 ngày trước

Predicting future physical states stands as a central challenge in embodied intelligence. Existing video world models leave 3D dynamics implicit, action-based formulations remain tied to specific robot configurations, and 3D motion prediction models use flow-only supervision, pro…

🅲 hạng C
20
Near-Zero Token Embeddings Silently Stall Training in Gated Sequence Models
📄 Paper HF Daily Papers 17 ngày trước

Among 291 surveyed open-weight language model repositories (280 distinct embedding tables), 52.6% of untied gated and hybrid checkpoints (20/38) contain reachable input embeddings with norms below 10^{-6}, compared with 0% of weight-tied models. Such near-zero rows can arise duri…

🅲 hạng C
20
RoboTryOn: Synthesizing Dexterous Robot Demonstrations from Human Videos via Virtual Try-On
📄 Paper HF Daily Papers 17 ngày trước

Robot foundation models have shown remarkable performance in robot control, yet adapting them to a new task still relies heavily on teleoperated demonstrations. Egocentric human videos offer a scalable substitute, but overlaying rendered robots on human videos or retargeting huma…

🅲 hạng C
20
Dory1.0-4B: A Looped Normalized-GPT
📄 Paper HF Daily Papers 18 ngày trước

We introduce Dory1.0-4B, a 4-billion-parameter looped (latent-recurrent) language model with elastic depth: a single checkpoint serves different inference compute budgets. Dory1.0-4B is built on a normalized Transformer (nGPT) backbone with interleaved sliding-window and full-att…

🅲 hạng C
20
Reducing Representational Interference in Retrieval Embeddings with Trainable Prefix Tokens
📄 Paper HF Daily Papers 15 ngày trước

Embedding models compress variable-length text into fixed-size vector representations, but this compression can entangle semantic content with residual token-level information, such as artifacts arising from attention sink behavior. We propose a simple method that prepends traina…

🅲 hạng C
20
Is Your Robot's Memory Just a Clock? Temporal Shortcuts in Memory-Augmented VLA Policies
📄 Paper HF Daily Papers 10 ngày trước

Memory-augmented vision-language-action (VLA) policies are meant to remember what happened. Many of them, however, fill memory by time: they sample frames uniformly across the episode and stamp each frame with its time step. Because time passes even when nothing happens, such a m…

🅲 hạng C
20
Are Causal Neuron Circuits Unique and Transferable?
📄 Paper HF Daily Papers 10 ngày trước

Interpretability demonstrations often show that intervening on a group of neurons changes a model's answer. A recent example removes an 18-neuron group in Qwen3-32B and the model answers "Simba" instead of "Scar" to a two-hop Lion King riddle. Such a demonstration establishes tha…

🅲 hạng C