Overfitting

Token-Set Choice Confounds POPE: A Systematic Audit of Yes/No Extraction in VLM Hallucination Evaluation

We set out to reproduce a recent hallucination-mitigation result on the POPE benchmark. The LLaVA-1.5-7B baseline we measured, 82.21 F1 on the adversarial split, sat 1.21 to 8.67 points above the adversarial baselines we planned to compare against (81.00, 79.30, and 73.54). We looked at the evaluation code and found a readout choice that moves F1 on its own. A two-token logit readout decides yes versus no by comparing two vocabulary IDs, 3582 ("yes") and 1217 ("no"). Acros…

I Vibe-Coded My Way Out of Angular Upgrades

Do you also hate upgrading Angular versions? Because I do. I maintain a few open-source projects, and...

LLMs reward expertise

Supplementary data: “Is metaphony (and sound change) sensitive to grammatical category? The case of verbs in the dialects of the Lausberg area" (Phonology)

This study examines the extent to which metaphony is sensitive to grammatical category in three varieties of the Lausberg dialect area of Southern Italy. Results from acoustic analyses show that metaphony, expressed through raising or diphthongisation of mid stem vowels triggered by high suffix vowels, is present in nouns and adjectives, but significantly reduced or absent in verbs. In certain cases, verbs display paradigmatic levelling, with metaphonic patterns extending even to forms lacking a triggering environment. These findings suggest that grammatical category can influence sound cha…

Significant-Gravitas/AutoGPT

AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

Is RTFx a good metric?

The performance of the transcription task for ASR models is evaluated using RTFx, which is the audio length divided by the time the model takes to transcribe the audio. The problem is that some ASR models (like Whisper) generate their output tokens in a autoregressive (token-by-token) manner, which means that the ti…

RoboValue: A Fine-Grained Sim-and-Real Benchmark for Unified Evaluation of Robotic Value Models

Robotic value models assess task execution from visual observations and instructions, providing feedback for policy learning and execution monitoring. However, existing evaluations cover only selected aspects of execution understanding, leaving the reliability of value feedback insufficiently tested. We introduce RoboValue, a fine-grained sim-and-real benchmark for unified evaluation of robotic value models. Its framework evaluates four complementary capability dimensions …

AI Avatar v25 (Avatar Run, Voxel Avatar, Muscle Brain)💪

Intro AI Avatar is a free app where your VRoid (VRM) avatar cheers you with all its...

LLMs are eroding my software engineering career and I don't know what to do

DS.01.10_Pancistrome of SeqOccIn Lines_2

Dataset 01.10 contains genomic tracks for transcription factor (TF) foot prints in hybrids of the SeqOccIn lines with Oh43 and allele-specific TF abundance counts at all SNPs covered by TF binding accourding to peak calling with MACS3. Additionally, p-values from a binomial test are provided to idicate whether there is allele-specific binding at the respective SNP. This part of the submission contains the track data for TF binding.

Word embedding

In natural language processing, a word embedding is a representation of a word. The embedding is used in text analysis. Typically, the representation is a real-valued vector that encodes the meaning of the word in such a way that the words that are closer in the vector space are expected to be similar in meaning. Word embeddings can be obtained using language modeling and feature learning techniques, where words or phrases from the vocabulary are mapped to vectors of real numbers.

huggingface/transformers

🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

What options exist for making a generative image model reproduce real typefaces in generated images?

What options exist for making a generative image model reproduce real typefaces in generated images, without needing actual font files (.ttf/.otf/etc)? Like retreival-augmented generation, training datasets using OCR'ed text and annotated typefaces, etc.

MemT-1.0: Towards Memory-Native Evolution

Memory-native evolution requires making memory an intrinsic capability of foundation models, enabling them to continuously accumulate and utilize information. To this end, we introduce MemT-1.0, a new architecture of memory foundation models that internalize memory as part of model computation. Building on Metis, MemT-1.0 employs Gated DeltaNet-2 with a stable token-level memory update mechanism within each turn. We further introduce Mixture-of-Storage to scale native memo…

Your Browser Can Do WHAT?! 5 AI Features That Actually Work

Are AI models running directly in the browser finally good enough to use in production? Let's find...

A small number of samples can poison LLMs of any size

Essays on Financial Machine Learning: Nonlinearity, Predictability and Data-Theory Integration

Gradient descent

Gradient descent is a method for unconstrained mathematical optimization. It is a first-order iterative algorithm for minimizing a differentiable multivariate function.

morluto/rea

Reverse engineer anything with agents, from app behavior down to native binaries.

What is the best way to train a model to distinguish between truncated and non-truncated objects from an image?

At the moment I am working on a vehicle counting & classification project. For a specific part in the project I need to get back only the completely visible vehicles from my input data (images). I am wondering if this could be done (more) automatically in the following way: zoom in such that only approximately o…

ActionRoPE: action as coordinates for spatially consistent control in planar-camera game world models

Game world models have made rapid progress toward real-time interactive generation. Yet action control is still predominantly implemented as feature conditioning, leaving the model to learn from training data how each action should move the visual world. This formulation is poorly matched to a broad class of planar-camera worlds, such as top-down and isometric games. In these worlds, player motion induces a geometrically well-defined planar camera displacement. We observe …

AI Fatigue? Yea I know.

I don't know how to start this off really haha. Though to be fair, everyone is talking about AI...

LLM Inevitabilism

Comparative assessment of AGAMOUS function in the Brassicaceae

Embedding (machine learning)

In machine learning, embedding is a representation learning technique that maps complex, high-dimensional data into a lower-dimensional vector space of numerical vectors.

langgenius/dify

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

NGBoost multi-step ahead probabilistic forecasting for 96 time steps

-1 I am building a probabilistic energy forecasting model using NGBoost (NGBRegressor) to predict household residual power (residual_power_w) at 15-minute resolution. My goal is to forecast the full next 24 hours (96 steps) with uncertainty estimates (prediction intervals, CRPS) at each step. The problem: NGBoost on…

Where is Physical RSI capability from? An Observational Scaling Law of LLM-as-policy

In this paper, we systematically investigate the mechanisms, scaling behavior, and generalization of physical Recursive self-improvement. We introduce PhysRSI-eval to quantify physical RSI through a five-axis metric: First-Attempt Performance, Efficiency, Success, Budget and Improvement from failure. Different VLMs are evaluated on 18 manipulation tasks under a fixed interaction budget. The analysis of experimental results motivates the Physical Barrel Hypothesis: differen…

One Weekend: 3,000 Open-Source AI Projects

Every October for the past twelve years, developers around the world have taken part in...

Show HN: Clippy – 90s UI for local LLMs

APPLICATION OF MACHINE LEARNING FOR MASS CYTOMETRY DATA ANALYSIS OF CHRONIC LYMPHOCYTIC LEUKAEMIA

Mass cytometry is a powerful technique for quantifying intracellular and membrane proteins at single-cell resolution. However, the vast amount of data it generates requires advanced analytical methods to extract meaningful biological insights. While machine learning (ML) has emerged as a key analytical tool in many fields, its application in mass cytometry remains underexplored. In this study, ML was employed to analyse mass cytometric data from primary Chronic Lymphocytic Leukaemia (CLL) samples, focusing on batch effect correction, classification, and marker association analysis. An ML-ba…

Computer vision

Computer vision tasks include methods for acquiring, processing, analyzing, and understanding digital images, and extraction of high-dimensional data from the real world in order to produce numerical or symbolic information, e.g. in the form of decisions. "Understanding" in this context signifies the transformation of visual images into descriptions of the world that make sense to thought processes and can elicit appropriate action. This image understanding can be seen as the disentangling of symbolic information from image data using models constructed with the aid of geometry, physics, statistics, and learning theory.

langchain-ai/langchain

The agent engineering platform.

Has artificial consciousness been studied in terms of intrinsic needs and limited agency?

In a sense, we humans are "doomed to think, doomed to desire, and doomed to want to act upon those desires when circumstances permit." We cannot simply escape our needs or limitations. Perhaps this is also what gives rise to our instinct for self-preservation and our sense of continuity as individuals with…

Kronecker Delta Rule for Long Context Linear Attention

We propose KronLSTM, a linear attention architecture based on the Kronecker Delta Rule, a novel update rule for associative matrix memories. KronLSTM extends the xLSTM family with a memory-revision mechanism that corrects targeted associations while modifying the memory only along a single direction in matrix space. This minimally invasive memory revision reduces interference with previously stored information compared with delta-rule updates, making it particularly well s…

Super-Intelligent Yes-Men: Are We Training AI to Ignore the Truth?

This is a submission for the Kaggle Benchmarking Challenge What I Benchmarked A published...

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL

Mathematical modelling of epidemics, evolution and social interaction

Natural language processing

Natural language processing (NLP) is the processing of natural language information by a computer. NLP is a subfield of computer science and is closely associated with artificial intelligence. NLP is also related to information retrieval, knowledge representation, computational linguistics, and linguistics more broadly.

firecrawl/firecrawl

Supercharge your AI agents with data from the web and beyond. Building the library for superintelligence. 🔥

How can we measure when open-source LLMs are "good enough" to replace proprietary frontier models like Claude/ChatGPT?

Frontier models from OpenAI, Anthropic, etc. keep improving, but the Chinese open-source/open-weight models are catching up. Open models could become good enough for most consumers and business applications, even if frontier models remain technically superior. Already DeepSeek is logging 3x as many tokens as Antroph…

SpecMRL: Redesign Matryoshka Representation Learning within Parametric Spectrum

Matryoshka Representation Learning (MRL) is designed to support flexible representation sizes by jointly optimizing nested prefixes. While effective, this design lacks an explicit constraint to ensure that the leading portion provides a suitable compact encoding. To address this, we propose SpecMRL, which reorganizes the feature space to enable more effective prefix readout. Specifically, we extend MRL with a parameter-derived spectral basis organized with feature importan…

AI Got Better While I Was Away. Software Didn't.

I haven't written here in a while. Not because I ran out of opinions. That would be concerning. I...

Things we learned about LLMs in 2024

Combating Data and Target Shifts in Visual Tasks

The rapid advancement of deep learning models for visual tasks has led to significant progress in many domains. However, a key challenge remains: ensuring that models can generalize effectively to unseen samples or novel classes, especially in real-world scenarios where training data-target pairs are often limited or unavailable, and test data may exhibit shifts in both data and targets. This thesis addresses the problem of shifts across data and targets, proposing novel methods to improve model generalization and transferability in complex settings. The thesis identifies the different shif…

Lexical analysis

Lexical tokenization is conversion of a text into meaningful lexical tokens belonging to categories defined by a "lexer" program. In case of a natural language, those categories include nouns, verbs, adjectives, punctuations etc. In case of a programming language, the categories include identifiers, operators, grouping symbols, data types and language keywords.

f/prompts.chat

f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.

How does parameter scaling work with generative image models?

How parameter scaling works with generative image models? Do generative image models require a large amount of parameters (e.g. between 20 and 40 billion parameters) to properly render detailed images and text? This applies to diffusion-based, autoregressive (AR) and hybrid (diffusion+AR) generative image models.

Text-to-Image Generative Modeling Without Language Pretraining

Modern text-to-image (T2I) systems use internet-scale language pretraining for text-conditioned image generation. We ask whether language pretraining is necessary for strong T2I generation. We introduce Halcyon, a family of early-fusion multimodal autoregressive models pretrained entirely from scratch on GPIC, a dataset of 100M image-text pairs. Halcyon-0.52B outperforms LlamaGen-XL/8 (48.0 vs. 63.3 gFD) with 18.2times less pretraining compute. Further, Halcyon-1.6B outper…

I got Jev to zero mistakes. I'm still using Flash-Lite.

Jev is a brilliant decision model. Gemini Flash-Lite is the model nobody talks about, and it’s fast,...

Learning to Reason with LLMs

Depth of Processing, learner aptitude, and the acquisition of L2 English grammatical structures

Depth of Processing (DoP) is central to several models of instructed second language acquisition (ISLA) (Leow, 2015). While DoP has been extensively examined in vocabulary learning (e.g., Laufer & Hulstijn, 2001), its role in the acquisition of grammatical structures, particularly through oral and audio-visual modalities remains comparatively underexplored. Drawing on the Levels of Processing framework (Craik & Lockhart, 1972), this study investigated whether task-induced processing depth influences the acquisition of third conditionals and comparatives, and whether language aptitude (LLAMA…

Large language model

A large language model (LLM) is an AI model trained on a vast amount of text for natural language processing tasks, especially language generation. LLMs can typically generate, summarize, translate, and analyze text in many contexts. They are the basis for many modern chatbots, such as ChatGPT, Claude, Gemini, Grok, Qwen and DeepSeek.

affaan-m/ECC

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

How was $m ≥ 8log(\frac{7|H|}{6})$ derived here?

Section 5.1 of Understanding Machine Learning: From Theory to Algorithms 2014 by Shai Shalev-Shwartz and Shai Ben-David states: With probability of at least $\frac{1}{7}$ over the choice of $S ∼ D^m$ we have that $L_D(A(S)) ≥ \frac{1}{8}$ . This theorem states that for every learner, there exists a task on which it …

HandsOn: A Large-Scale Simulation Benchmark for Generalizable Dexterous Manipulation

General-purpose manipulation requires reliable behavior across embodiments, objects, and task demands. We introduce HandsOn, a composable simulation platform and benchmark for cross-embodiment dexterous manipulation. It supports five robotic arms, six dexterous hands, and approximately 200 rigid, articulated, and deformable assets in randomized scenes. The benchmark pairs 18,000 curated human demonstrations with 66 human-verified tasks: Core-50 supports demonstration-based…

To Retry or Not to Retry? That Is the Question.

This is a submission for the Kaggle Benchmarking Challenge Those who read my articles know that a...

Llm.c – LLM training in simple, pure C/CUDA

Processed proteomics data and downstream analysis results for

This dataset contains processed protein-level quantification tables, differential-protein analysis results, PCA and clustering outputs, and downstream GO, GOSlim, KEGG and PPI analyses generated for this study. The dataset contains processed proteomics results and associated analysis reports. Vendor-native LC–MS/MS raw files and spectrum-level search outputs are not included.

Recurrent neural network

In artificial neural networks, recurrent neural networks (RNNs) are designed for processing sequential data, such as text, speech, and time series, where the order of elements is important. Unlike feedforward neural networks, which process inputs independently, RNNs utilize recurrent connections, where the output of a neuron at one time step is fed back as input to the network at the next time step. This enables RNNs to capture temporal dependencies and patterns within sequences.

ollama/ollama

Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.

Is LLM context just an extended prompt?

Does pasting the contents from skills files and project working files into a prompt, in front of the question (making the prompt very long), have the same effect as configuring the IDE to include those resources as "context" for the LLM? Does asking a second question in the same chat session have the same …

Vision-Language-Delta-Action: Learning Compact Deltas as Actionable Transition Interfaces for Robotic Manipulation

Robotic manipulation is not merely observation-to-action generation, but the realization of task-specified state changes in the physical world. Existing vision-language-action (VLA) methods directly learn observation-to-action mappings, imposing a substantial learning burden due to the gap between semantic understanding and action execution. Although world-action models (WAMs) improve control through future prediction, modeling complete future states wastes capacity on sta…

I Think We're Forgetting How to Be Bored

You know the moment. You're waiting for the kettle, or the elevator, or a page to load, or a friend...

The Era of 1-bit LLMs: ternary parameters for cost-effective computing

Processed CITE-seq data, spatial niche analysis pipeline and analysis code for conventional dendritic cells in mouse and human lung

Bayesian inference

Bayesian inference is a method of statistical inference in which Bayes' theorem is used to calculate a probability of a hypothesis, given prior evidence, and update it as more information becomes available. Fundamentally, Bayesian inference uses a prior distribution to estimate posterior probabilities. Bayesian inference is an important technique in statistics, and especially in mathematical statistics. Bayesian updating is particularly important in the dynamic analysis of a sequence of data. Bayesian inference has found application in a wide range of activities, including science, engineering, philosophy, medicine, sport, psychology, and law. In the philosophy of decision theory, Bayesian inference is closely related to subjective probability, often called "Bayesian probability".

open-webui/open-webui

User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

AdaExploreBench: Benchmarking LLM Agents for Proactive Exploration and Test-time Adaptation in Evolving Environments

Large Language Model (LLM) agents are increasingly deployed in unfamiliar and evolving environments, requiring them to proactively explore, understand, and adapt to environmental dynamics. However, existing benchmarks rarely test whether agents can proactively discover how an environment works and adapt as it evolves. To address this gap, we introduce AdaExploreBench, a benchmark for proactive exploration and test-time adaptation in evolving environments. By instantiating …

LLM Visualization

Decision responsiveness and the legitimacy of public agencies data set

The data were collected in relation to the following goals: The research was designed to identify some of the key factors driving assessments of the legitimacy of decision-making agencies. The research was based on a set of (vignette-based) experiments that allowed the researchers to identify the effects on legitimacy of various aspects of decision-making responsiveness. Participants (N=438) were drawn from Prolific Academic, and completed an online survey fielded on Qualtrics. Assignment to treatment groups was randomised.

Convolutional neural network

A convolutional neural network (CNN) is a type of feedforward neural network that learns features via filter optimization. This type of deep learning network has been applied to process and make predictions from many different types of data including text, images and audio. CNNs are the de-facto standard in deep learning-based approaches to computer vision and image processing, and have only recently been replaced—in some cases—by newer architectures such as the transformer.

DietrichGebert/ponytail

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

Scaling Parameter-Efficient Fine-Tuning for Brain Encoding

Brain encoding models aim to predict neural responses to stimuli. The most widely used approach fits readouts from frozen pretrained backbones. Using neural data to train large models from scratch or fully fine-tune them after task-pretraining has proven difficult because neural data is costly and limited. We show that parameter-efficient fine-tuning (PEFT) enables us to effectively adapt very large pretrained backbones for brain encoding with a small number of trainable p…

Llamafile lets you distribute and run LLMs with a single file

PAGAD: Projective Adversarial Generator for Image Anomaly Detection

Generative adversarial network

A generative adversarial network (GAN) is a class of machine learning frameworks and a prominent framework for approaching generative AI. The concept was initially developed by Ian Goodfellow and his colleagues at Université de Montréal in June 2014. In a GAN, two neural networks compete with each other in the form of a zero-sum game, where one agent's gain is another agent's loss.

Holi-World: Building Holistic World Intelligence from Human-Ego Experience

Human egocentric experience offers a rich foundation for learning how the world is organized and how it changes through interaction. Realizing this potential at scale requires turning growing volumes of recordings into learning resources that connect scene structure, object semantics, and interaction histories across viewpoints and time, while reducing dependence on costly manual annotation. We introduce HOLI-WORLD, an integrated ecosystem connecting experience capture, wo…

The evaluation of kinetic and static perimetry practice in neuro-ophthalmology

Reinforcement learning

In machine learning and optimal control, reinforcement learning (RL) is concerned with how an intelligent agent should take actions in a dynamic environment in order to maximize a reward signal. Reinforcement learning is one of the three basic machine learning paradigms, alongside supervised learning and unsupervised learning.

Learning from Fleet Experience: A Collective Learning Paradigm for Robots

Experience generated during deployment offers an opportunity for robots to improve policy reliability through learning from their own executions. However, learning through human corrections requires teleoperators to monitor execution and intervene when policies fail, a demand that is difficult to meet as robot fleets grow. We introduce collective experience learning (CEL), a learning paradigm in which deployed robots share their execution experience to improve a common pol…

Competing Learning in Hybrid Architectures

Many modern language models have adopted hybrid architectures that incorporate both linear-time recurrence and softmax attention. Modern recurrence layers provide a highly expressive mixing mechanism while ensuring high throughput by compressing contextual information into a fixed-size state, but underperform softmax attention in contextual information retrieval. In this paper, we reveal a Competing Learning phenomenon where different kinds of mixers compete to learn the s…

InfiniDream: Open-Ended Photorealistic 3D Worlds Through Procedural Generation

Procedural generation plays a key role in synthetic data creation and game development, offering greater controllability, editability, and diversity than static mesh representations. However, VLM-based code generation often produces relatively simple 3D assets, while sophisticated photorealistic procedural systems require extensive manual development. Existing VLM-guided procedural methods remain largely constrained by predefined factories and APIs, limiting their ability …