Universal Textual Teaching for LLMs
arXiv:2610.12114v1 Announce Type: new Abstract: Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student par
Knowledge catalogue
arXiv:2610.12114v1 Announce Type: new Abstract: Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student par
arXiv:2610.12281v1 Announce Type: cross Abstract: Over 90% of disease-associated variants from genome-wide association studies fall in noncoding regulatory regions, yet their functional interpretation
arXiv:2610.12161v1 Announce Type: cross Abstract: Why does removing reference-policy KL regularization sometimes improve group policy optimization? This motivates studying how reference-policy informa
arXiv:2610.11750v1 Announce Type: new Abstract: Tree-based speculative decoding verifies multiple draft continuations in one target-model pass, but finite trees built from draft scores face a fundamen
arXiv:2610.10805v1 Announce Type: new Abstract: As AI systems increasingly interact with people and make decisions about them, understanding human interpretations becomes an important part of developi
arXiv:2610.11247v1 Announce Type: cross Abstract: On-policy distillation (OPD) enables effective capability transfer between language models, yet the mechanisms underlying its failures are not fully u
arXiv:2610.10447v1 Announce Type: new Abstract: Reinforcement Learning (RL) from outcome rewards suffers from sparse supervision, particularly on difficult, long-horizon tasks where successful traject
arXiv:2509.19571v2 Announce Type: replace-cross Abstract: Designing or learning robot policies that generalize zero-shot across a range of language instructions and objects is a core problem in roboti
arXiv:2610.09661v1 Announce Type: new Abstract: Automatic Short Answer Scoring (ASAS) is central to NLP for Education. However, openly available benchmarks remain scarce, and existing datasets largely
arXiv:2610.09446v1 Announce Type: new Abstract: Large language models (LLMs) should abstain from scientific multiple-choice questions when no option is valid, but frequent abstention alone does not de
arXiv:2610.09337v1 Announce Type: cross Abstract: Policy-based reinforcement learning (RL) approaches have produced promising results for autonomous cyber defense; however, they are sample-inefficient
arXiv:2610.09026v1 Announce Type: cross Abstract: We present BEACON-SP, an ontology-grounded Graph Retrieval-Augmented Generation (GraphRAG) framework for clinician-facing decision support in behavior
arXiv:2610.09170v1 Announce Type: new Abstract: Autoregressive action-token policies such as vision-language-action models require action tokenizers to translate discrete token sequences into precise
arXiv:2610.09679v1 Announce Type: new Abstract: Adaptive rollout methods for group-relative reinforcement learning typically allocate a fixed per-update budget across prompts. We instead study how to
arXiv:2610.09541v1 Announce Type: cross Abstract: Signals that predict whether a chain-of-thought (CoT) trace is correct are compared by AUC, but deploying one requires a threshold with a guarantee. W
arXiv:2610.09219v1 Announce Type: cross Abstract: Direct Preference Optimization (DPO) treats all constraint violations equally: a 1 budget overshoot and a 1,000 overshoot induce the same training sig
arXiv:2610.08833v1 Announce Type: new Abstract: Masked diffusion language models (MDLMs) decode by repeatedly committing tokens to masked positions, but these commitments are usually irreversible. A t
arXiv:2610.09552v1 Announce Type: cross Abstract: Watermarking enables language model providers to identify text generated by their models. However, its desired properties can conflict (ie~stronger wa
arXiv:2610.08807v1 Announce Type: new Abstract: Precision pesticide spraying is essential for optimizing application efficiency and ensuring uniform chemical distribution. Spraying performance is infl
arXiv:2610.09174v1 Announce Type: new Abstract: Reliable and interpretable trajectory prediction is critical for cooperative and autonomous driving in complex and uncertain environments. This paper in
arXiv:2610.09358v1 Announce Type: new Abstract: Robots are increasingly expected to provide personalized services in everyday environments. To do so, they must ground natural-language commands such as
arXiv:2610.09597v1 Announce Type: new Abstract: Asynchronous RL accelerates large language model post-training by decoupling rollout generation from optimization, but trains on stale trajectories. Exi
arXiv:2610.10178v1 Announce Type: new Abstract: Vision-Language-Action models are designed to generalise across environments and task descriptions, raising the question of whether their action generat
arXiv:2610.09993v1 Announce Type: new Abstract: Spatial defect signatures on wafer bin maps (WBMs) trace yield loss to specific process faults, yet supervised classifiers recognize only the defect typ
arXiv:2610.10533v1 Announce Type: new Abstract: Conditional memory architectures such as DeepSeek Engram use input n-grams to look up learned embeddings, expanding the capacity of large language model
arXiv:2610.09372v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) layers replace the feed-forward block of a Transformer with E expert networks, and each token is routed to k of these experts.
arXiv:2610.10390v1 Announce Type: new Abstract: Vision-language-action~(VLA) models have emerged as a promising paradigm for autonomous driving. However, existing VLA models still suffer from a fundam
arXiv:2610.09202v1 Announce Type: cross Abstract: Extreme low-bit LLM compression is most challenging when weights, activations, and KV caches are quantized together: their distributions differ, and q
arXiv:2610.09412v1 Announce Type: cross Abstract: Retrieval diversification is widely available in retrieval-augmented generation (RAG) frameworks, yet prior studies disagree on whether it improves re
arXiv:2610.09958v1 Announce Type: new Abstract: Training can drive a transformer into a rank collapse: all token representations point in one direction, and learning stops. In a related collapse of at
This table to help anyone looking to build a budget Data Center homelab. I copied the bulk of value based, mid level, decent speed results GPUs and feed it to AI or SI and here are the recommended res
arXiv:2610.08959v1 Announce Type: new Abstract: On-policy distillation post-trains large language model agents by supplying dense, step-level guidance from a teacher policy when the reinforcement-lear
UT-based small models, despite being trained from scratch on these tasks without internet-scale pre-training, consistently outperform most standard Transformer-based Large Language models (LLMs) by a
arXiv:2610.10203v1 Announce Type: new Abstract: Model organisms of alignment-relevant behaviors (e.g., backdoors, sycophancy, spurious correlations) have emerged as a key tool for evaluating whitebox
arXiv:2610.09940v1 Announce Type: cross Abstract: Joint-embedding predictive architectures (JEPAs) predict masked or future observations in representation space, offering a natural source of predictiv
Last week I was called into a meeting with OpenAI’s head of safety and told they no longer trust me. A security guard took my badge and walked me out of the building. Then I learned my colleagues @bal
arXiv:2610.06647v2 Announce Type: replace Abstract: Reinforcement learning has greatly advanced the capabilities of large language models, but its memory demands remain a barrier to broader adoption.
arXiv:2610.10114v1 Announce Type: new Abstract: The architectural design of Large Language Models (LLMs) is shifting from traditional full-attention-only models to hybrid models, which combine differe
arXiv:2610.09145v1 Announce Type: cross Abstract: We revisit a standard accepted practice in the continuous diffusion language model literature of fixing conditioning prompt tokens clean during traini
arXiv:2610.10381v1 Announce Type: new Abstract: Looped Transformers improve parameter efficiency by repeatedly applying shared Transformer blocks over multiple recurrent loops, increasing computationa
arXiv:2509.18597v3 Announce Type: replace Abstract: Large language models (LLMs) can translate natural-language instructions for robotic manipulation into executable code, but ambiguity, noisy generat
arXiv:2610.10310v1 Announce Type: new Abstract: Recursive self-improvement requires carrying accepted changes into later improvement cycles, while studying agent-proposed changes also requires substan
arXiv:2610.09294v1 Announce Type: cross Abstract: Rapid progress in AI agents has brought growing attention to agent safety, with extensive evaluation focused on digital environments. As agents move i
I spent the past year independently building Slow Vale, an LLM-driven life simulation. The Chinese server now has 800+ AI residents sharing one continuously running city. This is an engineering write-
arXiv:2610.10444v1 Announce Type: cross Abstract: Much knowledge work produces new deliverables from files a workspace already holds, and LLM agents are beginning to take such work over. Through direc
arXiv:2509.00141v2 Announce Type: replace-cross Abstract: Statutory corpora and judicial decisions are growing faster than legal professionals can read them, while individual judgments often exceed th
arXiv:2610.09611v1 Announce Type: new Abstract: Large neural networks often acquire capabilities that small models fail to learn. Does this stem from large models learning more representative features
arXiv:2610.09032v1 Announce Type: new Abstract: Perceiving structured shapes, such as human faces, from pixels is an inherently ambiguous task in real-world conditions. Yet, shape inference is largely
arXiv:2610.09034v1 Announce Type: cross Abstract: Multi-agent path planning (MAPP) in continuous environments often relies on roadmaps to balance safety and search efficiency. However, traditional roa
arXiv:2602.08234v2 Announce Type: replace Abstract: Large Language Model (LLM) agents have shown stunning results in complex tasks, yet they often operate in isolation, failing to learn from past expe
Hi all, having searched google and asked several AIs without success, maybe s.o. can help. I have 2 x dgx spark in a cluster. deepseek-flash is running successful. Now I want to test qwen3.8-flash-nex
arXiv:2610.09074v1 Announce Type: new Abstract: Emerging long-horizon agentic tasks require repeated model calls, worsening the inference cost of already-costly language models. While narrow agentic t
arXiv:2610.08820v1 Announce Type: new Abstract: Large language model-based multi-agent systems improve complex problem solving through collaboration, while latent communication directly transmits mode
Hey everyone, Jovan from UkisAI (Swift Qwen) here! For those who don't know us, UkisAI is a small lab making tiny frontier LLMs, tools and datasets (+doing it open-source!). I'm one of the guys runnin
arXiv:2610.09048v1 Announce Type: new Abstract: Building a world model requires a state representation useful for planning and decision-making---potentially over tasks unknown at training time. In the
arXiv:2610.10411v1 Announce Type: cross Abstract: Speculative decoding accelerates large language model inference by using a low-cost draft model to propose tokens that the full-size target model veri
arXiv:2610.10183v1 Announce Type: new Abstract: Long video understanding increasingly relies on external memory to organize massive visual streams into compact representations. However, most memory-ba
arXiv:2610.08828v1 Announce Type: new Abstract: Small-data adaptation can improve speech detection while degrading speaker attribution. We study this discrepancy in a released streaming diarizer adapt
arXiv:2610.09471v1 Announce Type: new Abstract: Learning systems adapt quickly inside a familiar family of models. The harder step comes earlier: deciding, from observations that could be noise, an ex
arXiv:2610.10508v1 Announce Type: new Abstract: Prompted embedding models have recently received increasing attention, particularly for retrieval, where detailed retrieval instructions are provided as