Large Language Models (LLMs) are increasingly trained to abstain from answering questions they are unsure about. However, this ability is often misused: in real-world applications, input prompts sometimes contain uncertainty elements, and driven by this, LLMs are inclined to abstain even on problems they are capable of solving. We argue that LLM abstention is not only an expression of genuine uncertainty; it is also an artifact that can be largely influenced by prompts. We name this phenomenon *Abstention Inflation*. We add "Unknown" as an extra option for LLMs to choose from; experiments show serious accuracy drops on True/False Questions (TFQs). Replacing "Unknown" with an unrelated random word produces an identical effect. We argue that LLMs are trained to imitate the surface pattern of abstention, rather than to express genuine uncertainty. Based on ten experiments, we support four claims that form a progressive argument: **(C1)** *Abstention Inflation* is triggered by the structural presence of an extra option, not by genuine uncertainty; **(C2)** further, it makes the model deny it can answer even when it can; **(C3)** at the representation level, this manifests as a later-layer output override; **(C4)** finally, this bias is stable and emerges through instruction tuning, rather than stochastic noise.

Large language models (LLMs) have become increasingly used for various tasks, often coupled with chain-of-thought (CoT) prompting to boost accuracy. Recent work has shown that high label-prediction accuracy does not guarantee correct intermediate reasoning, and the causes of *reasoning flaws* vary from sample to sample, yet existing remedies either focus on a single domain or assume that one flaw type applies uniformly across samples. A simple mitigation method is to provide the model with the correct answer, but we show that this yields no consistent improvement in reasoning quality. This indicates that the problem cannot be fixed by LLMs’ awareness of answers, and must instead be addressed through the structure of reasoning. Motivated by this, we propose **CRAFT** (Consensus Reasoning-knowledge-graph Aggregation for Flaw-aware Trace synthesis), which aggregates the consensus components shared across multiple candidate reasoning traces to synthesize improved ones. **CRAFT** consistently improves label-prediction accuracy on both logical and mathematical reasoning benchmarks, outperforming most baselines, while its post-processed traces achieve higher quality under fine-grained benchmark evaluation.

Logical reasoning is a core capability for large language models (LLMs), yet existing benchmarks that rely solely on final-answer accuracy fail to capture the quality of the reasoning process. To address this, we introduce FineLogic, a fine-grained evaluation framework that assesses logical reasoning across three dimensions: overall accuracy, stepwise soundness, and representation-level probing. Leveraging this framework, we conduct a comprehensive study on how different supervision formats in fine-tuning shape reasoning abilities. We fine-tune LLMs on four supervision styles—one in natural language and three symbolic variants—and find a key trade-off: natural language supervision excels at generalization to out-of-distribution and long-chain problems, whereas symbolic supervision is superior at instilling structurally sound, atomic reasoning steps. Furthermore, our probing analysis indicates that fine-tuning primarily refines the model’s step-by-step generation process, rather than improving its ability to converge on an answer early. Together, our framework and analysis provide a more rigorous lens for evaluating and improving logical reasoning in LLMs. The code is available at https://github.com/YujunZhou/FineLogic.