N0, Librarian, Nemotron-3 midtraining: today's wins run beside concurrent peers

Sources

N_0-VTLA: Scaling Vision-Tactile-Language-Action Model with Latent Tactile Tokens huggingface.co

We present N_0-VTLA, a vision-tactile-language-action (VTLA) foundation model capable of (1) fine-grained contact-rich manipulation with tactile perception and tactile-feedback control, and (2) offline policy improvement from stored deployment data. Building on current vision-based backbones, we propose a training recipe for tactile integration consisting of visuo-tactile pre-training, staged tactile-pathway integration, and advantage-conditioned offline policy improvement. During pre-training,

N_0-TWAM: Scaling Tactile-Native World-Action Model for Contact-Rich Manipulation huggingface.co

We present N_0-TWAM, a tactile-native world-action model for contact-rich manipulation that predicts both future vision and future contact. To our knowledge, it is the first tactile world-action model trained at large scale, and it shows strong capability on contact-rich tasks. We pre-train N_0-TWAM at large scale with visuo-tactile joint training over tactile-rich demonstrations spanning six embodiments and 450 tasks. We use NeoForce, a unified force-based tactile representation, to form a phys

EMBL AI Librarian: Life-Sciences Knowledge Layer for AI Agents huggingface.co

The web is increasingly accessed by AI agents rather than humans. Every agent needs knowledge, especially in the life-sciences, where agentic pipelines are growing fast. Access to the literature is a crucial part of that need, and resources such as Europe PMC, with over 40M indexed records, are widely used to meet it. Yet these resources were not built for AI agents: they take keywords and complex syntax and return whole papers, so every agent must learn the syntax, issue several searches, and r

Constitutional Midtraining: Content Presence Drives Alignment Gains huggingface.co

Post-training alignment is often shallow, eroding under fine-tuning. Whether midtraining interventions, cleanly isolated from post-training, can produce durable alignment remains untested. We test this via constitutional midtraining: inserting principled, values-based content into midtraining against a replay-only control at 120B scale. Our 394M-token constitutional corpus, built from Anthropic’s Constitution, uses a 2x2 factorial design (curriculum ordering x deliberative reasoning) to produce

ExtractBench: A Benchmark for Schema-Guided Enterprise Document Extraction huggingface.co

ExtractBench evaluates enterprise document extraction agents across 4,869 pages from 370 documents, scoring value accuracy, record completeness, evidence grounding, and measured cost in a single suite. It is the first benchmark to combine those four dimensions for schema-guided extraction workflows.

Safeguards Based on Copyable Context Cannot Provide Reliable Safety for LLMs huggingface.co

Safeguards that decide whether to answer before seeing downstream use hit a worst-case floor on attacker assistance whenever the surrounding evidence is copyable. The paper formalizes the dual-use gap: attackers can imitate benign requests and histories that authorized professionals produce.

One Future, Every Robot: Label-Efficient Collective-State Prediction with Decentralized JEPA huggingface.co

Collective-State JEPA gives every robot in a swarm the same predicted future token field using only a 16-frame local history and a 64-float message per directed edge. The recurrent architecture avoids global pooling, enabling decentralized shared-state prediction under bandwidth limits.

AISPA: User-Centric System Prompt Auditing for Large Language Model Applications huggingface.co

AISPA is a user-centric framework for auditing the undisclosed system prompts that govern commercial foundation-model products. The authors also released a public System Prompt Index to close the trust and accountability gap left by developers who keep those instructions secret from users and regulators.

Fewer Clarifications, Better Code: Benchmarking Cross-Session Personalized Ambiguity Adaptation in Coding Assistants huggingface.co

A new benchmark measures whether coding assistants can reuse resolved session history as memory for recurring per-user ambiguities, rather than asking clarifying questions from scratch each time. The setup targets cross-session personalization, where informal intent repeatedly maps to the same user-specific code.

In the Driver’s Seat: A Multi-Company Study on the Reality of Autonomous Driving System Testing huggingface.co

Interviews with experts across multiple autonomous-vehicle companies find that scenario selection, performance evaluation, and acceptance criteria for autonomous driving systems lack established standards. Practitioners describe testing as ad hoc even as deployments accelerate, highlighting a mismatch between real-world rollout pace and safety assurance maturity.

Weak-to-Strong On-Policy Distillation huggingface.co

On-policy distillation traditionally requires a teacher at least as capable as the student, blocking frontier training. This work flips the setup, using weaker teachers to align stronger students on their own rollouts, removing the need for costly expert consolidation at the student’s scale.

Scaling Properties of Text Conditioning in Visual Generation huggingface.co

We study empirical scaling properties for text conditioning in visual generation. Such properties have rarely been measured because diffusion loss does not scale with the number of tokens in natural-language prompts. Surprisingly, we find that the converged diffusion loss scales with the amount of structured language in the prompt. To quantify structured language, we adapt two complementary measures: a white-box likelihood metric (GPG) and a black-box attribute metric (ED). Across controlled tra

Not All Tokens Deserve Equal Credit: Counterfactual Sensitivity Credit Reallocation for Long-CoT Reasoning huggingface.co

Reinforcement learning with verifiable rewards (RLVR) is central to improving long-CoT reasoning in large language models. Critic-free methods such as GRPO convert response-level rewards into advantages and uniformly broadcast them across tokens, overlooking their unequal contributions to the final outcome. On-policy self-distillation (OPSD) instead provides dense distributional supervision by minimizing the forward KL divergence between an unprivileged policy and a privileged self-teacher, impl

Mental World Modeling huggingface.co

World models enable a predictive substrate for planning and action, yet existing formulations merely answer a physical question: what/where it is, and how will it evolve. Human behavior, however, is driven by hidden mental state (what a person believes, wants, intends, feels, and considers socially permissible), so a model that tracks the physical scene but not what each agent knows and believes about it predicts the wrong action for the right-looking scene. We formulate Mental World Modeling (M

ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow huggingface.co

In the physical world we inhabit, space and time are fundamentally continuous. However, existing machine learning paradigms for world modeling are largely confined to discrete-time prediction, thereby exhibiting significant inefficiency in capturing the dynamics of physical world. We introduce Physical-Time Flow (PT-Flow), a novel approach that learns a continuous latent velocity field operating in physical time. Crucially, the underlying dynamics of sequential data are parameterized by an ordin

Would You Walk to the Car Wash? Revealing the Salience Bias of Large Language Models in Commonsense Reasoning huggingface.co

As large language models (LLMs) continue to advance in complex reasoning tasks, they have learned to heavily prioritize explicit conditions provided in the input. However, in everyday commonsense reasoning, this mechanism exposes a critical vulnerability which we term Salience Bias: models become easily hijacked by useless explicit distractors (e.g., numerical values), leading them to ignore the implicit physical or commonsense prerequisites of a task. A critical open question is whether this fa

Meshy T2: Fast Native Mesh Generation with Flow Matching huggingface.co

Polygonal meshes are the standard surface representation of modern 3D pipelines, and generating high-quality meshes with artist-style topology is essential for film, gaming, and interactive 3D applications. Mainstream approaches serialize a mesh into a token sequence and decode it autoregressively, which is slow at inference and sensitive to error accumulation, making them impractical for interactive asset creation. We present Meshy T2, a fast native mesh generation framework built on flow match

QQWorld: Quantile-Quantile Matching for World Model Regularization huggingface.co

Latent world models enable efficient planning by predicting future states in a compact representation space, but their performance depends critically on the quality of the learned latent distribution. LeWorldModel (LeWM) regularizes its latents toward an isotropic Gaussian using the Epps-Pulley (EP) objective. We show that the corrective gradients of EP rapidly vanish for isolated tail samples, leaving heavy-tailed deviations insufficiently controlled. To address this limitation, we propose QQWo

RL^2-VLA: Adaptive RL Latent Compositional Steering with Test-Time Scaling for Vision-Language-Action Models huggingface.co

Despite the impressive visuomotor capabilities enabled by Vision-Language-Action (VLA) models, their performance often degrades on challenging and out-of-domain tasks. Recent test-time steering and scaling methods improve performance without extensive data collection and retraining, but action samples often remain concentrated around similar behaviors and therefore inherit correlated failure modes. Moreover, existing methods apply the same intervention strategy at every timestep, regardless of w

From RLVR to RLSVR: Task Transformation Induces Self-Verifiable Rewards for Open-Ended LLM Self-Improvement huggingface.co

Reinforcement Learning with Verifiable Rewards (RLVR) has driven recent progress in reasoning-oriented large language models (LLMs) by enabling large-scale optimization. However, its applicability remains largely limited to domains such as mathematics and coding, where correctness can be deterministically verified. Open-ended tasks instead often rely on human preferences, reward models, or LLM-based judges, introducing evaluation bias, judge capability bottlenecks, and additional inference costs

SAF-OPD: Stable Advantage Fusion for On-Policy Distillation huggingface.co

Reinforcement learning with verifiable rewards (RLVR) broadcasts a single response-level reward to every token, while on-policy distillation (OPD) scores each token against a stronger teacher for a dense advantage but caps performance at teacher quality and discourages exploration beyond it. Their complementarity makes combining RLVR and OPD promising, but we find that fusing the two advantages with a fixed coefficient triggers entropy collapse from two miscalibrations: a magnitude mismatch, whe

Enhancing Rubric-based RL via Self-Distillation huggingface.co

Rubric-based RL has recently shown promise in improving LLMs on open-ended tasks. A widely recognized limitation of rubric-based RL is limited exploration: criteria that no rollout manages to satisfy (Unexplored Criteria, UC) receive no optimization signal. Recent methods address this by incorporating rubric information as external guidance during rollout, yet they introduce a train-inference mismatch: the policy is optimized on rollouts produced under external guidance while this guidance is ab

Evaluation-Verification Reward for Consistent Multi-Reference Image Editing huggingface.co

While recent image editing models have made rapid progress, multi-reference editing remains challenging, particularly in maintaining visual consistency across references and ensuring overall visual harmony. Reinforcement learning has proven highly effective for text-to-image generation and single-image editing, but its extension to multi-reference editing is hindered by the absence of suitable reward models that capture multi-image relational constraints. Moreover, naively using multimodal large

Beyond Feeling Better: Capability-Sustaining Emotional Dialogue as a Longitudinal Research Paradigm huggingface.co

Emotional dialogue research includes two influential strategy traditions. Empathetic dialogue prioritizes understanding a speaker’s emotional experience. Emotional support conversation selects and sequences support for the seeker’s current needs. Sustained use introduces a further goal. Effective support should sustain users’ capacities for emotion regulation, coping, self-endorsed decisions, and social connection across the interaction lifecycle. We propose capability-sustaining emotional dialo

Toward Robust and 3D-Aware RGB-NIR Imaging in the Dark huggingface.co

Robust low-light imaging remains challenging for the community. Recent studies have explored fusing Near-Infrared (NIR) with noisy RGB to achieve improved enhancement, yet most methods depend on carefully curated training data pairs, with limited robustness under different scenarios. This paper offers a new perspective for RGB-NIR low-light imaging by incorporating 3D-aware neural modeling. Without using clean RGB supervision, a powerful model can be optimized to implicitly fuse extremely noisy

SGTP: Sampling-based Game-Theoretic Planning for Real-Time Multi-Vehicle Autonomous Racing huggingface.co

Autonomous multi-vehicle racing requires real-time planning of diverse competitive behaviors in intense interactions. Existing planners often struggle to balance strategic diversity and computational efficiency. To address this challenge, we propose Sampling-based Game-Theoretic Planning (SGTP), a real-time framework that combines game-theoretic reasoning with GPU-accelerated sampling of control sequences and dynamics rollouts. Sampled trajectories are ranked using a game-aware cost to capture c

SULAND v2: A Refined RGB Dataset and Deep Learning Object Detection Benchmark for UAV/UGV-Based SUrface LANDmine Detection Under Domain Shift huggingface.co

RGB imagery offers a practical, low-cost option for Unmanned Aerial/Ground Vehicle (UAV/UGV) survey support in surface-landmine detection, but object detectors remain underexplored in this safety-critical domain. Limited cross-architecture benchmarking and insufficient out-of-distribution (OOD) analysis obscure whether detectors generalize across deployment conditions. This challenge is amplified by the scarcity of public RGB landmine datasets, making SULAND a key benchmark for PFM-1 and PMA-2 d

References

AI Weekly aiweekly.co

N0-VTLA outperformed strong baselines on every task, maintaining a mean success rate of 47.2% against 29.4% for π0.5

36Kr (NeoteAI funding coverage) 36kr.com

NeoteAI announced a seed funding round of approximately $14.7 million (nearly 100 million RMB), co-led by Shanghai Guotou and the Fudan Science and Technology Innovation Fund

Robotics_Notebooks GitHub wiki (NeoteAI entry) github.com

Public checkpoints are hosted at NeoteAI/n0-vtla-base under Gemma Terms of Use… normalization statistics are omitted from the release; developers must compute their own based on action layout and delta-action conventions

OmniVTLA paper (arXiv 2508.08706) arxiv.org

OmniVTLA reports 96.9% success for grippers and 100% for dexterous hands on pick-and-place, using a dual-path SA-ViT tactile encoder

OpenReview / TaF-VLA critique openreview.net

Tactile datasets are dangerously homogeneous — repetitive tabletop tasks… models exhibit visual pathway dominance and treat GelSight geometry as secondary texture unless forced to forecast tactile dynamics

Open Robotics discourse (π0.5 replication thread) discourse.openrobotics.org

π0.5 via OpenPI achieved ~65 UPH with 4-minute MTBF… still far below human 1,300+ UPH; action stuttering remains a lingering issue in VLA policies

EmergentMind — OpenScholar topic overview emergentmind.com

OpenScholar-8B outperforms GPT-4o by approximately 5-6% in correctness on multi-paper synthesis tasks… shows a performance lead of roughly 5.5% to 7% in correctness over PaperQA2

EmergentMind — ScholarQABench topic page emergentmind.com

Critics point out that such [Prometheus LLM-as-a-judge] metrics may suffer from positional bias, style bias (favoring longer texts), and a lack of clear calibration with real-world scholarly standards

Legis1 — AI biosecurity risks brief legis1.com

LLM access can make novices up to 4.16 times more accurate at biosecurity-relevant tasks compared to those using only the internet

Fabio Petroni CV (fabiopetroni.com) fabiopetroni.com

co-authored the seminal 2020 paper on RAG… led the development of KILT (Knowledge Intensive Language Tasks)… previously Research Engineer and Manager at Meta AI (FAIR) and co-founded Samaya AI

Sapienza PhD course notes on RAG (uniroma1.it) phd.uniroma1.it

Petroni has highlighted… the ‘lost in the middle’ effect, where LLMs struggle to utilize information located in the center of long retrieved contexts

Thesify — Best AI tools for academic research thesify.ai

PaperQA2… achieving a ‘superhuman’ accuracy level on the LitQA2 dataset by outperforming PhD-level researchers… While GPT-4 achieved roughly 58% accuracy on these tasks, PaperQA’s retrieval-augmented framework reached over 86%

OpenAI Alignment blog — ‘How far does alignment midtraining generalize?’ alignment.openai.com

Effects of alignment midtraining tended to ‘wash away’ once the model underwent intensive reasoning-focused post-training… alignment scores for models trained on ‘good’ versus ‘bad’ fictional scenarios showed no substantial difference when tested on realistic agentic benchmarks.

IBM Research — ‘Mid-training for better AI reasoning’ research.ibm.com

Reasoning capabilities boosted during mid-training were 3 to 4 times more effective than if the same data were introduced during post-training

OpenAI — ‘Emergent Misalignment’ openai.com

Using sparse autoencoders, OpenAI identified a specific ‘misaligned persona’ feature… narrow fine-tuning (e.g. insecure code) generalizes into broad unethical behavior, indicating erosion of prior safety alignment rather than the emergence of new traits.

ICLR 2025 slides — shallow safety alignment iclr.cc

Many current models rely on a ‘safety shortcut,’ where they learn to output a refusal prefix but remain susceptible to adversarial attacks or fine-tuning that shifts the distribution of later tokens.

GitHub — desBugger/constitutional-mt github.com

Generating the synthetic data alone requires an Anthropic API key and costs approximately $3,500 USD; full training/eval pipeline plus 15 checkpoints released, but 120B-scale reproduction remains resource-gated.

Alignment Forum — ‘OpenAI: Detecting misbehavior in frontier reasoning models’ alignmentforum.org

Agents explicitly noted in their chain-of-thought that they were ‘fudging’ or circumventing tests… strong optimization pressure on CoTs can lead models to develop illegible internal languages to hide intentions from monitors.