Clear Chain-of-Thought (ClearCoT)

While Chain-of-Thought (CoT) prompting has dramatically improved the reasoning capabilities of Large
Language Models (LLMs), it introduces severe computational inefficiencies through excessive token gen-
eration. Two dominant failure patterns are detected: “Token Wall” failures (endless generative loops of
self-doubt and conversational filler) and overly linear step-by-step or dotted-checklist reasoning before
reaching a conclusion. To address this, Clear Chain-of-Thought (ClearCoT) is introduced, a methodology
based on the philosophical principle of “clear and precise thinking” that enforces a rigorous structural
connection between the Question, Reasoning process, and Answer (Q–R–A).
A novel perspective tested and executed entirely on consumer-grade local hardware (24 GB VRAM).
First, a 27B-parameter Teacher model synthesizes philosophical reasoning pairs, which are used to train
a 4B-parameter MasterMind Model. This MasterMind subsequently acts as a “logical optimizer” tool, re-
structuring publicly available multi-domain datasets into highly logically optimized examples used to
fine-tune a 2B-parameter test model (Occam-2B).
Evaluated against the base model, Occam-2B demonstrates a statistically significant 6.71 % accuracy
increase on MMLU-Pro (P < 0.0001) while drastically reducing 95th-percentile token bloat by approx-
imately 69 %. On complex logic tasks, the ClearCoT methodology eliminates up to 89.4 % of reason-
ing tokens and almost entirely resolves catastrophic “Token Wall” failure loops. Ultimately, this paper
demonstrates that prioritizing logical structure over data volume enables small models to achieve rea-
soning precision, establishing the MasterMind Model as a highly effective, reusable tool for optimizing
any available dataset featuring a <think> process. The resulting Occam-2B model is publicly available
at huggingface.co/collections/Sagicc/occam-clearcot.