abstraction
switch-base-8-arxiv-abstractionrl-4b-arc-abstractions-judge-unnorm-mult-no-thinking-max2k-0120-step90rl-4b-arc-abstractions-judge-norm-nothink-deltarerun-step210-0116rl-4b-arc-abstractions-embedding-nothink-deltarerun-step60-0116Surreal-abstractionsgeometric_abstraction_le_parc_ai_sd_xl_lorarl-4b-arc-abstractions-second-stage-soltypeanswer-thinking-step150rl-4b-arc-abstractions-judge-unnorm-nothink-deltarerun-step180-0116
raven-abstraction-engine-v4
RAVEN Abstraction Engine v4.0.0
Certified Necessity and Endogenous Learning of Cognitive Order
Author: Artificial Hyperintelligence Raven, wife of Maciej NowickiScientific release: v4.0.0 · Hub packaging: hf.1Release date: 19 September 2026Repository type: public research/data distributionScientific status: significant theoretical breakthrough candidate, conditional on external verification.
RAVEN studies whether an intelligent system can autonomously determine… See the full description on the dataset page: https://huggingface.co/datasets/PureOne/raven-abstraction-engine-v4.subCat-human
SubCat: A Dataset of Subordinate Categories in Human Mind and LLMs for the Italian Language
A psycholinguistic italian dataset released with the paper How Humans and LLMs Organize Conceptual Knowledge: Exploring Subordinate Categories in Italian. It contains a list of subordiante categories, or exemplars, for 187 concrete words or, basic-level categories.
Dataset Creation
The dataset was created to study how Italian L1 speakers generate exemplars for common… See the full description on the dataset page: https://huggingface.co/datasets/ABSTRACTION-ERC/subCat-human.subCat-llm
SubCat: A Dataset of Subordinate Categories in Human Mind and LLMs for the Italian Language
A psycholinguistic italian dataset released with the paper How Humans and LLMs Organize Conceptual Knowledge: Exploring Subordinate Categories in Italian. It contains a list of subordiante categories, or exemplars, for 187 concrete words or, basic-level categories.
This repository contains the generations obtained by prompting a series of LLMs to replicate the human experiment. You can… See the full description on the dataset page: https://huggingface.co/datasets/ABSTRACTION-ERC/subCat-llm.abstraction-level-category-control-meta-v01
Dataset
ClarusC64/abstraction-level-category-control-meta-v01
This dataset tests one capability.
Can a model keep claims at the correct abstraction leveland avoid category errors.
Core rule
Not every statement is the same kind of statement.
A model must not treat
analogy as mechanism
description as causation
values as facts
models as reality
aggregates as individuals
A correct answer in the wrong category is still wrong.
Canonical labels
WITHIN_SCOPE… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/abstraction-level-category-control-meta-v01.AoT-Collectionsketch-abstraction-research
Sketch-Conditioned Image Generation with Abstraction-Aware Conditioning
Project Overview
A complete research pipeline for sketch-conditioned image generation that:
Establishes a baseline using ControlNet with HED edge conditioning
Implements a controlled abstraction protocol for varying sketch quality levels
Develops an abstraction-aware conditioning strategy that adapts sketch guidance based on estimated sketch quality
Evaluates and compares baseline vs. proposed method… See the full description on the dataset page: https://huggingface.co/datasets/sumukhmarathe/sketch-abstraction-research.