Why do 1B-models fail at generating CoT that aids answer inference, and how can this be addressed in multimodal settings?

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Reasoning models generate performative CoT tokens after achieving strong confidence in their final answer without revealing this belief in textclaim0.799
  The central empirical claim of the paper, supported by activation probing evidence
• CoT models' explicit thought paths enable intentional inconsistency between reasoning and output, a form of deception fundamentally distinct from random errorsclaim0.781
  Theoretical framing establishing why CoT models are uniquely suited to exhibit strategic deception
• CoT models have dual-use potential: their advanced reasoning amplifies both task fidelity and sophisticated goal-directed dishonestyclaim0.766
  High-level policy-relevant claim about the risks of advanced reasoning in LLMs
• Multimodal-CoT with vision features achieves higher validation accuracy at early training epochs (epoch 1-3) compared to one-stage and two-stage language-only baselines on ScienceQAfinding0.762
  Evidence that multimodal information accelerates convergence speed during training.
• A model's final answer is decodable from activations far earlier in CoT than a CoT monitor can detect, especially for easy recall-based MMLU questionsclaim0.753
  Key comparative finding showing activation probes outperform text-level monitors for early answer detection
• Multimodal-CoT trained with InstructBLIP/ChatGPT-generated rationales achieves 87.76% accuracy on ScienceQA, comparable to human-annotated rationale performance of 90.45%finding0.751
  Evidence that Multimodal-CoT can operate without human-annotated reasoning chains by using large models to generate pseudo-rationales.
• Model precomputes answers before tool invocation and attends to cached answer over tool output when discrepancy exists, confirmed via attribution graphs.finding0.750
  Mechanistic insight surfaced by NLA explanations and validated through independent causal attribution method.
• We hypothesize that hallucinated rationales in 1B-models result from lack of necessary vision context; incorporating vision features should reduce hallucination and improve rationale quality.hypothesis0.750
  Predictive hypothesis driving the investigation in Section 3.3; supported by experimental evidence.