The Cognitive Companion: A Lightweight Parallel Monitoring Architecture for Detecting and Recovering from Reasoning Degradation in LLM Agents
Summary: arXiv:2604.13759v1 Announce Type: new
Abstract: Large language model (LLM) agents on multi-step tasks suffer reasoning degradation, looping, drift, and stuck states, at rates up to 30% on hard tasks. Current solutions include hard step limits (abrupt) or LLM-as-judge monitoring (10-15% overhead per step). This paper introduces the Cognitive Companion, a parallel monitoring architecture with two implementations: an LLM-based Companion and a novel zero-overhead Probe-based Companion.
We report a three-batch feasibility study centered on Gemma 4 E4B, with an additional exploratory small-model analysis on Qwen 2.5 1.5B and Llama 3.2 1B. In our experiments, the LLM-based Companion reduced repetition on loop-prone tasks by 52-62% with approximately 11% overhead. The Probe-based Companion, trained on hidden states from layer 28, showed a mean effect size of +0.471 at zero measured inference overhead; its strongest probe result achieved cross-validated AUROC 0.840 on a small proxy-labeled dataset.
Key Findings
A key empirical finding is that companion benefit appears task-type dependent:
- Companions are most helpful on loop-prone and open-ended tasks.
- Effects are neutral or negative on more structured tasks.
Our small-model experiments also suggest a possible scale boundary: companions did not improve the measured quality proxy on 1B-1.5B models, even when interventions fired. Overall, the paper should be read as a feasibility study rather than a definitive validation.
Conclusion
The results provide encouraging evidence that sub-token monitoring may be useful, identify task-type sensitivity as a practical design constraint, and motivate selective companion activation as a promising direction for future work. Further studies will be necessary to explore the implications of these findings and to refine the Cognitive Companion architecture for broader applications in LLM agent performance enhancement.