paper-with-me

홈 › Papers

Ask Twice, Look Twice: Prompt Echoing Resolves the Question-First Paradox in Vision-Language Models

2026-07-17 · Rakshanda Hassan Abhinandan, John Galeotti, Deva Ramanan, Gautam Rajendrakumar Gare arxiv

Where should the question go in a vision-language model (VLM) prompt: before the image or after it? Intuition says before: knowing what is asked should tell the model where to look. Yet across visual question answering benchmarks, question-first prompting consistently underperforms the image-first ordering recommended for frontier VLMs, a phenomenon we term the question-first paradox. We trace the paradox to a conflict between two stages of VLM computation. Logit-lens and attention probes show the intuition is half right: a question placed before the image genuinely steers perception, moving image patch representations toward question-relevant concepts. The failure lies downstream. Stranded behind hundreds of image tokens, the question is barely attended by the answer token, which instead commits to image-driven (often wrong) answers; a causal attention knockout confirms that the answer reads the question only when the question follows the image. The diagnosis yields a training-free fix: question echoing, restating the question on both sides of the image so that one copy steers perception while the other is read out at answer time. The same division of labor appears in a fifty-year-old finding on human ``adjunct questions'', where repeating a question before and after a passage aids comprehension more than either position alone. Echoing the image as well brings further gains, restoring the whole-image view a causal decoder otherwise loses. The paradox holds across five open VLMs, costing up to 17.5 group-accuracy points. Echoed prompts close it and surpass the best single-pass ordering on NaturalBench, POPE, Winoground, and open-ended VQAv2, by up to 19 Winoground group-accuracy points, with no training, fine-tuning, or architecture change. The paradox reveals a trade-off between steering perception and preserving question access; echoing resolves it through prompt design alone.

📄 PDF Abstract BibTeX arXiv:2607.15565

Code (0)

등록된 구현이 없습니다.

Tasks

Visual Question Answering

Similar Papers 제목 키워드 기반

TWICE: Two-Clock, Two-Window Learning for Long-Horizon Conversion Prediction in Online Advertising

2026-07-28 · Kaiyuan Li, Kun Wang, Zhongbo Wang, Teng Sha 외 arxiv

Long-horizon conversion prediction under delayed feedback creates a two-clock, two-window learning problem in online advertising. A short base observation window releases recent clicks on the click clock before their out…

ThinkTwice: Jointly Optimizing Large Language Models for Reasoning and Self-Refinement

2026-04-02 · Difan Jiao, Qianfeng Wen, Blair Yang, Zhenwei Tang 외 arxiv

We introduce ThinkTwice, a simple two-phase framework that jointly optimizes LLMs to solve reasoning problems and refine the answers, based on Group Relative Policy Optimization (GRPO). In each pair of training steps, Th…

Mathematical Reasoning

Do not be greedy, Think Twice: Sampling and Selection for Document-level Information Extraction

2026-01-26 · Mikel Zubillaga, Oscar Sainz, Oier Lopez de Lacalle, Eneko Agirre arxiv

Document-level Information Extraction (DocIE) aims to produce an output template with the entities, relations, and events of interest occurring in the given document. Standard practices include prompting decoder-only LLM…

Information Extraction

ReadTwice: Reading Very Large Documents with Memories

2021-05-10 · NAACL 2021 4 · Yury Zemlyanskiy, Joshua Ainslie, Michiel de Jong, Philip Pham 외

Knowledge-intensive tasks such as question answering often require assimilating information from different sections of large inputs such as books or article collections. We propose ReadTwice, a simple and effective techn…

Question Answering

Look Twice Before You Answer: Memory-Space Visual Retracing for Hallucination Mitigation in Multimodal Large Language Models

2024-10-04 · Xin Zou, Yizhou Wang, Yibo Yan, Yuanhuiyi Lyu 외

Despite their impressive capabilities, multimodal large language models (MLLMs) are prone to hallucinations, i.e., the generated content that is nonsensical or unfaithful to input sources. Unlike in LLMs, hallucinations …

DecoderHallucination