Research map / Learning theory and algorithms / Combinatorial and search optimization
KitBit: A New AI Model for Solving Intelligence Tests and Numerical Series
Víctor Corsino, José Manuel Gilpérez, Luis Herrera
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| continue search for all solutions instead of stopping at first, 1-level constancy (ablation) | KitBit · Literature series (67) | solved (%): 88.1 -> 89.6 | Improved |
| require 2-level constancy instead of 1-level (stop at solution) (ablation) | KitBit · Literature series (67) | solved (%): 88.1 -> 73.1 | Hurt |
| require 2-level constancy instead of 1-level (ablation) | KitBit · IQ test series (90) | solved (%): 97.8 -> 97.8 | No clear effect |
| increase number of terms of some series (ablation) | KitBit · Literature series (67) | solved (%): 73.1 -> 91.0 | Improved |
| increase state tree depth from 1 to 4 (ablation) | KitBit · OEIS | share of resolved sequences by depth (%): 23.4%, 21.4%, 28.2% and 27.0% | Improved |
| KitBit instead of Ragni and Klein ANN | KitBit · OEIS | solved series: 26,951 -> 87,514 | Improved |
| evaluate on selected OEIS subset instead of entire database | KitBit · OEIS | solved (%): 25.6 -> 36.0 | Improved |
Closest papers by meaning
- OlympicArena: Benchmarking Multi-discipline Cognitive Reasoning for Superintelligent AI (NeurIPS 2024 · similarity 0.83)
- Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles (NeurIPS 2025 · similarity 0.83)
- Creativity or Brute Force? Using Brainteasers as a Window into the Problem-Solving Abilities of Large Language Models (NeurIPS 2025 · similarity 0.83)
- MLR-Bench: Evaluating AI Agents on Open-Ended Machine Learning Research (NeurIPS 2025 · similarity 0.82)
- SciCode: A Research Coding Benchmark Curated by Scientists (NeurIPS 2024 · similarity 0.82)
- PUZZLES: A Benchmark for Neural Algorithmic Reasoning (NeurIPS 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
