WMIP: A Working-Memory-Inspired Programmable Probe of Cumulative State Tracking in Large Language Models
cs.AI, cs.CL
Submitted: 2026-03-28
Updated: 2026-10-30
Code: https://github.com/dengzhe-hou/WMF-AM
Terminology
Sources
- AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using Agents
- M3-BENCH: Process-Aware Evaluation of LLM Agents' Social Behaviors in Mixed-Motive Games
- Let's Verify Step by Step
- Solving math word problems with process- and outcome-based feedback
- MemGPT: Towards LLMs as Operating Systems
- On the Failure of Latent State Persistence in Large Language Models
- The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
- Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting
- Measuring Faithfulness in Chain-of-Thought Reasoning
- Fragile Thoughts: How Large Language Models Handle Chain-of-Thought Perturbations
- Robust Answers, Fragile Logic: Probing the Decoupling Hypothesis in LLM Reasoning
- Exploring State Tracking Capabilities of Large Language Models
- (How) Do Language Models Track State?
- Strong Memory, Weak Control: An Empirical Study of Executive Functioning in LLMs
- A Neuropsychologically Grounded Evaluation of LLM Cognitive Abilities
- Unable to Forget: Proactive Interference Reveals Working Memory Limits in LLMs Beyond Context Length
- Intelligence Degradation in Long-Context LLMs: Critical Threshold Determination via Natural Length Distribution Analysis
- Cognitive Load Limits in Large Language Models: Benchmarking Multi-Hop Reasoning
- Show Your Work: Scratchpads for Intermediate Computation with Language Models
- Qwen2.5 Technical Report
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents