LBLuis F. Becerra-Monsalve

Luis F. Becerra-Monsalve

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Luis F. Becerra-Monsalve has a total of 2 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Automatic Functional Characterization of Chains of Thought Generated by Reasoning Models

    by Luis F. Becerra-Monsalve, German Sanchez-Torres, John W. Branch-Bedoya

    Reasoning models generate chains of thought (gCoT) that externalize deliberations, calculations, verifications, corrections and alternatives, and are today a principal source of information about their behavior. The literature examines their faithfulness to internal computation, answer correctness, length and step quality, or describes their structure with step-level graphs and trees; multi-step operations and their relations remain less characterized. This work proposes describing that organization through a functional profile obtained with an automatic, language-model-supported method that represents each chain as a graph, identifies its episodes and summarizes their relations. Validity and reliability were tested: on a control corpus the method detected all explicit deliberative episodes and 71% of implicit ones; on real chains, group means of 25 chains proved reliable. In 200 chains from four models on shared medical and legal problems, organization differed between models and was similar across domains; verifications rarely refuted (8%) and alternatives were rarely adopted (2%). The results suggest that the profile captures each model’s features: it allowed the generating model to be identified in 67% of chains, versus 49% with correctness, length and domain, and added no detectable information about correctness. Profiles describe the observable chain organization at the group level, not the models’ internal process.

  • Open Access

    Multi-Dimensional Evaluation of Auto-Generated Chain-of-Thought Traces in Reasoning Models

    by Luis F. Becerra-Monsalve, German Sanchez-Torres, John W. Branch-Bedoya

    Automatically generated chains-of-thought (gCoTs) have become common as large language models adopt deliberative behaviors. Prior work emphasizes fidelity to internal processes, leaving explanatory properties underexplored. Our central hypothesis is that these traces, produced by highly capable reasoning models, are not arbitrary by-products of decoding but exhibit stable and practically valuable textual properties beyond answer fidelity. We apply a multidimensional text-evaluation framework that quantifies four axes—structural coherence, logical–factual consistency, linguistic clarity, and coverage/informativeness—that are standard dimensions for assessing textual quality, and use it to evaluate five reasoning models on the GSM8K arithmetic word-problem benchmark (~1.3 k–1.4 k items) with reproducible, normalized metrics. Logical verification shows near-ceiling self-consistency, measured by the Aggregate Consistency Score (ACS ≈ 0.95–1.00), and high final-answer entailment, measured by Final Answer Soundness (FAS0 ≈ 0.85–1.00); when sound, justifications are compact, with Justification Set Size (JSS ≈ 0.51–0.57) and moderate redundancy, measured by the Redundant Constraint Ratio (RCR ≈ 0.62–0.70). Results also show consistent coherence and clarity; from gCoT to answer implication is stricter than from question to gCoT support, indicating chains anchored to the prompt. We find no systematic trade-off between clarity and informativeness (within-model slopes ≈ 0). In addition to these automatic and logic-based metrics, we include an exploratory expert rating of a subset (four raters; 50 items × five models) to contextualize model differences; these human judgments are not intended to support dataset-wide generalization. Overall, gCoTs display explanatory value beyond fidelity, primarily supported by the automated and logic-based analyses, motivating hybrid evaluation (automatic + exploratory human) to map convergence/divergence zones for user-facing applications.