LLM Digest

Story

arxiv_llm_reliability · Nov 2, 2026 · paper

Source brief

MIRA: Reinforcing Multimodal Reasoning via Deceptive Contextual Augmentation

arxiv.orgNov 2, 2026
original source linked

In brief

While Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced multimodal reasoning, existing frameworks suffer from a critical "perception-reasoning gap." Due to an over-reliance on seemingly relevant visual tokens, even minor perceptual perturbations can

Read the original at arxiv.org →Open in live feed Read that day’s brief

MIRA: Reinforcing Multimodal Reasoning via Deceptive Contextual Augmentation

Earlier in this thread 4 items

Latent Reasoning in Continuous Space for Unified Multimodal Models

Counterfactual Pairwise Augmentation Improves Small Pointwise LLM Judges

Judging Before Propagation: Task-Graph Contextual Integrity for Multi-Agent Privacy

AgentDecept: Benchmarking Deceptive Behavior in Tool-Using LLM Agents under Pressure