CoTu at EXACT 2026: Neuro-Symbolic Reasoning for Transparent Educational QA
On the map: LLM reasoning → Reasoning benchmarks and evaluation
Abstract (arXiv)
Transparent educational question answering asks for answers that are not only correct but explainable, and doing so with small models rules out the reasoning power of the largest proprietary systems. The EXACT 2026 competition poses this problem concretely: open-weight language models of at most 8B parameters, self-hosted, with a natural-language explanation for every answer. It pairs two tasks: logical reasoning over university regulations, and multi-step physics problem solving. We describe the system that team cotu developed to address both, a neuro-symbolic Program-of-Thought pipeline in which a 4B backbone writes a program rather than stating an answer directly: for regulation queries it emits a Z3 encoding whose entailment verdict grounds the deduction, and for physics it emits numerical Python, both wrapped in a shared self-correction loop and a unified explained-JSON output. Answer-type routing, distillation-based task fine-tuning, and a latency-aware serving stack -- SGLang with speculative decoding -- keep the system within the 60-second per-query limit. The system achieved a perfect score on the physics task in both automated selection rounds and obtained the highest final-round technical score of any team -- 13.44/15, combining automated answer evaluation with expert-judged reasoning depth -- with the equally weighted presentation score included, cotu placed 3rd overall. Grounding answers in a symbolic solver yields correct, verifiable deductions at the 4B scale, and the residual difficulty lies in premise selection rather than the deduction itself.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 13 of its 30 references that are in the map. First posted on arXiv 16 Jul 2026.
Builds on
13 of 30 references.
Closest ancestors · 1
Same direction, newest first
Other references · 12
- 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
- 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
- 2023A Survey of Large Language Modelscited by 2,537
- 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
- 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 803
- 2024Large Language Models: A Surveycited by 506
- 2023Scaling Relationship on Learning Mathematical Reasoning with Large Language Modelscited by 417
- 2023Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Modelscited by 325
- 2023A Rank Stabilization Scaling Factor for Fine-Tuning with LoRAcited by 231
- 2026DFlash: Block Diffusion for Flash Speculative Decodingcited by 120
- 2025ACECODER: Acing Coder RL via Automated Test-Case Synthesiscited by 89
- 2024SelfCodeAlign: Self-Alignment for Code Generationcited by 61
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026A Verifier-Guided Explainable Reasoning Framework with Gold-Anchored QLoRA, Task-Aware Mixture-of-Experts, and Group-Relative RLVR
- 2024FLARE: Faithful Logic-Aided Reasoning and Exploration
- 2026Moxia: A Trust-First Neuro-Symbolic Execution Architecture for Self-Explaining Mathematical Reasoning
- 2025Bridging LLMs and Symbolic Reasoning in Educational QA Systems: Insights from the XAI Challenge at IJCNN 2025
- 2023Language Models can be Logical Solvers
- 2024Improve Student's Reasoning Generalizability through Cascading Decomposed CoTs Distillation
- 2026NeSy-RAG: Neuro-Symbolic RAG for Explainable Question Answering
- 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering