instance_id stringlengths 19 23 | item_id stringclasses 221
values | origin stringclasses 2
values | image stringclasses 221
values | image_path stringclasses 221
values | task stringclasses 5
values | task_name stringclasses 5
values | question stringclasses 171
values | answer stringclasses 84
values | answer_aliases listlengths 0 0 | candidates listlengths 0 4 | level stringclasses 5
values | level_mode stringclasses 5
values | explanation_reference stringclasses 221
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
human_seed_0001__H0 | human_seed_0001 | human_seed_figure | images/human/叹为观止.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 叹为观止 | [] | [] | {"answer": "叹为观止", "visible_cues": [{"id": "P1", "content": "人物在叹气"}, {"id": "P2", "content": "观音形象制止叹气者"}], "bridge_units": [{"id": "B1", "image_concept": "叹气", "acceptable_intermediate_concepts": [], "target_idiom_part": "叹", "gold_bridge_path": ["叹气", "叹"], "evidence": ["P1"], "required": true}, {"id": "B2", "image_... | |||
human_seed_0001__H1 | human_seed_0001 | human_seed_figure | images/human/叹为观止.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 叹为观止 | [] | [] | {"answer": "叹为观止", "visible_cues": [{"id": "P1", "content": "人物在叹气"}, {"id": "P2", "content": "观音形象制止叹气者"}], "bridge_units": [{"id": "B1", "image_concept": "叹气", "acceptable_intermediate_concepts": [], "target_idiom_part": "叹", "gold_bridge_path": ["叹气", "叹"], "evidence": ["P1"], "required": true}, {"id": "B2", "image_... | |||
human_seed_0001__H4 | human_seed_0001 | human_seed_figure | images/human/叹为观止.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 叹为观止
B. 指鹿为马
C. 门可罗雀
D. 凿壁偷光
请只输出最可能的成语，不要输出选项字母，不要解释。 | 叹为观止 | [] | [
"叹为观止",
"指鹿为马",
"门可罗雀",
"凿壁偷光"
] | {"answer": "叹为观止", "visible_cues": [{"id": "P1", "content": "人物在叹气"}, {"id": "P2", "content": "观音形象制止叹气者"}], "bridge_units": [{"id": "B1", "image_concept": "叹气", "acceptable_intermediate_concepts": [], "target_idiom_part": "叹", "gold_bridge_path": ["叹气", "叹"], "evidence": ["P1"], "required": true}, {"id": "B2", "image_... | |||
human_seed_0001__E0 | human_seed_0001 | human_seed_figure | images/human/叹为观止.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 叹为观止 | [] | [] | {"answer": "叹为观止", "visible_cues": [{"id": "P1", "content": "人物在叹气"}, {"id": "P2", "content": "观音形象制止叹气者"}], "bridge_units": [{"id": "B1", "image_concept": "叹气", "acceptable_intermediate_concepts": [], "target_idiom_part": "叹", "gold_bridge_path": ["叹气", "叹"], "evidence": ["P1"], "required": true}, {"id": "B2", "image_... | |||
human_seed_0001__E1 | human_seed_0001 | human_seed_figure | images/human/叹为观止.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“叹为观止”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "叹为观止",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 叹为观止 | [] | [] | {"answer": "叹为观止", "visible_cues": [{"id": "P1", "content": "人物在叹气"}, {"id": "P2", "content": "观音形象制止叹气者"}], "bridge_units": [{"id": "B1", "image_concept": "叹气", "acceptable_intermediate_concepts": [], "target_idiom_part": "叹", "gold_bridge_path": ["叹气", "叹"], "evidence": ["P1"], "required": true}, {"id": "B2", "image_... | |||
human_seed_0002__H0 | human_seed_0002 | human_seed_figure | images/human/头头是道.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 头头是道 | [] | [] | {"answer": "头头是道", "visible_cues": [{"id": "P1", "content": "几个人认某个角色为老大"}, {"id": "P2", "content": "中间角色是道士形象"}], "bridge_units": [{"id": "B1", "image_concept": "老大", "acceptable_intermediate_concepts": ["头头"], "target_idiom_part": "头头", "gold_bridge_path": ["老大", "头头"], "evidence": ["P1"], "required": true}, {"id": "... | |||
human_seed_0002__H1 | human_seed_0002 | human_seed_figure | images/human/头头是道.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 头头是道 | [] | [] | {"answer": "头头是道", "visible_cues": [{"id": "P1", "content": "几个人认某个角色为老大"}, {"id": "P2", "content": "中间角色是道士形象"}], "bridge_units": [{"id": "B1", "image_concept": "老大", "acceptable_intermediate_concepts": ["头头"], "target_idiom_part": "头头", "gold_bridge_path": ["老大", "头头"], "evidence": ["P1"], "required": true}, {"id": "... | |||
human_seed_0002__H4 | human_seed_0002 | human_seed_figure | images/human/头头是道.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 分道扬镳
B. 头头是道
C. 背道而驰
D. 素不相识
请只输出最可能的成语，不要输出选项字母，不要解释。 | 头头是道 | [] | [
"分道扬镳",
"头头是道",
"背道而驰",
"素不相识"
] | {"answer": "头头是道", "visible_cues": [{"id": "P1", "content": "几个人认某个角色为老大"}, {"id": "P2", "content": "中间角色是道士形象"}], "bridge_units": [{"id": "B1", "image_concept": "老大", "acceptable_intermediate_concepts": ["头头"], "target_idiom_part": "头头", "gold_bridge_path": ["老大", "头头"], "evidence": ["P1"], "required": true}, {"id": "... | |||
human_seed_0002__E0 | human_seed_0002 | human_seed_figure | images/human/头头是道.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 头头是道 | [] | [] | {"answer": "头头是道", "visible_cues": [{"id": "P1", "content": "几个人认某个角色为老大"}, {"id": "P2", "content": "中间角色是道士形象"}], "bridge_units": [{"id": "B1", "image_concept": "老大", "acceptable_intermediate_concepts": ["头头"], "target_idiom_part": "头头", "gold_bridge_path": ["老大", "头头"], "evidence": ["P1"], "required": true}, {"id": "... | |||
human_seed_0002__E1 | human_seed_0002 | human_seed_figure | images/human/头头是道.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“头头是道”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "头头是道",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 头头是道 | [] | [] | {"answer": "头头是道", "visible_cues": [{"id": "P1", "content": "几个人认某个角色为老大"}, {"id": "P2", "content": "中间角色是道士形象"}], "bridge_units": [{"id": "B1", "image_concept": "老大", "acceptable_intermediate_concepts": ["头头"], "target_idiom_part": "头头", "gold_bridge_path": ["老大", "头头"], "evidence": ["P1"], "required": true}, {"id": "... | |||
human_seed_0003__H0 | human_seed_0003 | human_seed_figure | images/human/拖泥带水.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 拖泥带水 | [] | [] | {"answer": "拖泥带水", "visible_cues": [{"id": "P1", "content": "图中出现钢铁侠 / 托尼斯塔克"}, {"id": "P2", "content": "钢铁侠帮人带来一瓶水"}], "bridge_units": [{"id": "B1", "image_concept": "托尼斯塔克", "acceptable_intermediate_concepts": ["托尼", "Tony"], "target_idiom_part": "拖泥", "gold_bridge_path": ["托尼斯塔克", "托尼", "拖泥"], "evidence": ["P1"], "r... | |||
human_seed_0003__H1 | human_seed_0003 | human_seed_figure | images/human/拖泥带水.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 拖泥带水 | [] | [] | {"answer": "拖泥带水", "visible_cues": [{"id": "P1", "content": "图中出现钢铁侠 / 托尼斯塔克"}, {"id": "P2", "content": "钢铁侠帮人带来一瓶水"}], "bridge_units": [{"id": "B1", "image_concept": "托尼斯塔克", "acceptable_intermediate_concepts": ["托尼", "Tony"], "target_idiom_part": "拖泥", "gold_bridge_path": ["托尼斯塔克", "托尼", "拖泥"], "evidence": ["P1"], "r... | |||
human_seed_0003__H4 | human_seed_0003 | human_seed_figure | images/human/拖泥带水.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 泥古不化
B. 闻鸡起舞
C. 拖泥带水
D. 东施效颦
请只输出最可能的成语，不要输出选项字母，不要解释。 | 拖泥带水 | [] | [
"泥古不化",
"闻鸡起舞",
"拖泥带水",
"东施效颦"
] | {"answer": "拖泥带水", "visible_cues": [{"id": "P1", "content": "图中出现钢铁侠 / 托尼斯塔克"}, {"id": "P2", "content": "钢铁侠帮人带来一瓶水"}], "bridge_units": [{"id": "B1", "image_concept": "托尼斯塔克", "acceptable_intermediate_concepts": ["托尼", "Tony"], "target_idiom_part": "拖泥", "gold_bridge_path": ["托尼斯塔克", "托尼", "拖泥"], "evidence": ["P1"], "r... | |||
human_seed_0003__E0 | human_seed_0003 | human_seed_figure | images/human/拖泥带水.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 拖泥带水 | [] | [] | {"answer": "拖泥带水", "visible_cues": [{"id": "P1", "content": "图中出现钢铁侠 / 托尼斯塔克"}, {"id": "P2", "content": "钢铁侠帮人带来一瓶水"}], "bridge_units": [{"id": "B1", "image_concept": "托尼斯塔克", "acceptable_intermediate_concepts": ["托尼", "Tony"], "target_idiom_part": "拖泥", "gold_bridge_path": ["托尼斯塔克", "托尼", "拖泥"], "evidence": ["P1"], "r... | |||
human_seed_0003__E1 | human_seed_0003 | human_seed_figure | images/human/拖泥带水.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“拖泥带水”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "拖泥带水",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 拖泥带水 | [] | [] | {"answer": "拖泥带水", "visible_cues": [{"id": "P1", "content": "图中出现钢铁侠 / 托尼斯塔克"}, {"id": "P2", "content": "钢铁侠帮人带来一瓶水"}], "bridge_units": [{"id": "B1", "image_concept": "托尼斯塔克", "acceptable_intermediate_concepts": ["托尼", "Tony"], "target_idiom_part": "拖泥", "gold_bridge_path": ["托尼斯塔克", "托尼", "拖泥"], "evidence": ["P1"], "r... | |||
human_seed_0004__H0 | human_seed_0004 | human_seed_figure | images/human/纹丝不动.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 纹丝不动 | [] | [] | {"answer": "纹丝不动", "visible_cues": [{"id": "P1", "content": "图中有蚊子"}, {"id": "P2", "content": "图中有螺丝钉"}, {"id": "P3", "content": "蚊子和螺丝钉被枪指着而不敢动"}], "bridge_units": [{"id": "B1", "image_concept": "蚊子", "acceptable_intermediate_concepts": [], "target_idiom_part": "纹", "gold_bridge_path": ["蚊子", "纹"], "evidence": ["P1"],... | |||
human_seed_0004__H1 | human_seed_0004 | human_seed_figure | images/human/纹丝不动.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 纹丝不动 | [] | [] | {"answer": "纹丝不动", "visible_cues": [{"id": "P1", "content": "图中有蚊子"}, {"id": "P2", "content": "图中有螺丝钉"}, {"id": "P3", "content": "蚊子和螺丝钉被枪指着而不敢动"}], "bridge_units": [{"id": "B1", "image_concept": "蚊子", "acceptable_intermediate_concepts": [], "target_idiom_part": "纹", "gold_bridge_path": ["蚊子", "纹"], "evidence": ["P1"],... | |||
human_seed_0004__H4 | human_seed_0004 | human_seed_figure | images/human/纹丝不动.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 大可不必
B. 怒不可遏
C. 泥古不化
D. 纹丝不动
请只输出最可能的成语，不要输出选项字母，不要解释。 | 纹丝不动 | [] | [
"大可不必",
"怒不可遏",
"泥古不化",
"纹丝不动"
] | {"answer": "纹丝不动", "visible_cues": [{"id": "P1", "content": "图中有蚊子"}, {"id": "P2", "content": "图中有螺丝钉"}, {"id": "P3", "content": "蚊子和螺丝钉被枪指着而不敢动"}], "bridge_units": [{"id": "B1", "image_concept": "蚊子", "acceptable_intermediate_concepts": [], "target_idiom_part": "纹", "gold_bridge_path": ["蚊子", "纹"], "evidence": ["P1"],... | |||
human_seed_0004__E0 | human_seed_0004 | human_seed_figure | images/human/纹丝不动.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 纹丝不动 | [] | [] | {"answer": "纹丝不动", "visible_cues": [{"id": "P1", "content": "图中有蚊子"}, {"id": "P2", "content": "图中有螺丝钉"}, {"id": "P3", "content": "蚊子和螺丝钉被枪指着而不敢动"}], "bridge_units": [{"id": "B1", "image_concept": "蚊子", "acceptable_intermediate_concepts": [], "target_idiom_part": "纹", "gold_bridge_path": ["蚊子", "纹"], "evidence": ["P1"],... | |||
human_seed_0004__E1 | human_seed_0004 | human_seed_figure | images/human/纹丝不动.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“纹丝不动”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "纹丝不动",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 纹丝不动 | [] | [] | {"answer": "纹丝不动", "visible_cues": [{"id": "P1", "content": "图中有蚊子"}, {"id": "P2", "content": "图中有螺丝钉"}, {"id": "P3", "content": "蚊子和螺丝钉被枪指着而不敢动"}], "bridge_units": [{"id": "B1", "image_concept": "蚊子", "acceptable_intermediate_concepts": [], "target_idiom_part": "纹", "gold_bridge_path": ["蚊子", "纹"], "evidence": ["P1"],... | |||
human_seed_0005__H0 | human_seed_0005 | human_seed_figure | images/human/小鸟依人.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 小鸟依人 | [] | [] | {"answer": "小鸟依人", "visible_cues": [{"id": "P1", "content": "小鸟在给人看病"}, {"id": "P2", "content": "小鸟像医生一样开药并说明吃法"}, {"id": "P3", "content": "图中有人作为病人"}], "bridge_units": [{"id": "B1", "image_concept": "小鸟", "acceptable_intermediate_concepts": [], "target_idiom_part": "小鸟", "gold_bridge_path": ["小鸟", "小鸟"], "evidence": [... | |||
human_seed_0005__H1 | human_seed_0005 | human_seed_figure | images/human/小鸟依人.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 小鸟依人 | [] | [] | {"answer": "小鸟依人", "visible_cues": [{"id": "P1", "content": "小鸟在给人看病"}, {"id": "P2", "content": "小鸟像医生一样开药并说明吃法"}, {"id": "P3", "content": "图中有人作为病人"}], "bridge_units": [{"id": "B1", "image_concept": "小鸟", "acceptable_intermediate_concepts": [], "target_idiom_part": "小鸟", "gold_bridge_path": ["小鸟", "小鸟"], "evidence": [... | |||
human_seed_0005__H4 | human_seed_0005 | human_seed_figure | images/human/小鸟依人.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 路人皆知
B. 正人君子
C. 小鸟依人
D. 惊弓之鸟
请只输出最可能的成语，不要输出选项字母，不要解释。 | 小鸟依人 | [] | [
"路人皆知",
"正人君子",
"小鸟依人",
"惊弓之鸟"
] | {"answer": "小鸟依人", "visible_cues": [{"id": "P1", "content": "小鸟在给人看病"}, {"id": "P2", "content": "小鸟像医生一样开药并说明吃法"}, {"id": "P3", "content": "图中有人作为病人"}], "bridge_units": [{"id": "B1", "image_concept": "小鸟", "acceptable_intermediate_concepts": [], "target_idiom_part": "小鸟", "gold_bridge_path": ["小鸟", "小鸟"], "evidence": [... | |||
human_seed_0005__E0 | human_seed_0005 | human_seed_figure | images/human/小鸟依人.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 小鸟依人 | [] | [] | {"answer": "小鸟依人", "visible_cues": [{"id": "P1", "content": "小鸟在给人看病"}, {"id": "P2", "content": "小鸟像医生一样开药并说明吃法"}, {"id": "P3", "content": "图中有人作为病人"}], "bridge_units": [{"id": "B1", "image_concept": "小鸟", "acceptable_intermediate_concepts": [], "target_idiom_part": "小鸟", "gold_bridge_path": ["小鸟", "小鸟"], "evidence": [... | |||
human_seed_0005__E1 | human_seed_0005 | human_seed_figure | images/human/小鸟依人.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“小鸟依人”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "小鸟依人",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 小鸟依人 | [] | [] | {"answer": "小鸟依人", "visible_cues": [{"id": "P1", "content": "小鸟在给人看病"}, {"id": "P2", "content": "小鸟像医生一样开药并说明吃法"}, {"id": "P3", "content": "图中有人作为病人"}], "bridge_units": [{"id": "B1", "image_concept": "小鸟", "acceptable_intermediate_concepts": [], "target_idiom_part": "小鸟", "gold_bridge_path": ["小鸟", "小鸟"], "evidence": [... | |||
human_seed_0006__H0 | human_seed_0006 | human_seed_figure | images/human/孜孜不倦.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 孜孜不倦 | [] | [] | {"answer": "孜孜不倦", "visible_cues": [{"id": "P1", "content": "一个人很困"}, {"id": "P2", "content": "另一个人用水枪喷水"}, {"id": "P3", "content": "被喷水后变精神不困了"}], "bridge_units": [{"id": "B1", "image_concept": "喷水声", "acceptable_intermediate_concepts": ["滋滋"], "target_idiom_part": "孜孜", "gold_bridge_path": ["喷水", "滋滋", "孜孜"], "eviden... | |||
human_seed_0006__H1 | human_seed_0006 | human_seed_figure | images/human/孜孜不倦.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 孜孜不倦 | [] | [] | {"answer": "孜孜不倦", "visible_cues": [{"id": "P1", "content": "一个人很困"}, {"id": "P2", "content": "另一个人用水枪喷水"}, {"id": "P3", "content": "被喷水后变精神不困了"}], "bridge_units": [{"id": "B1", "image_concept": "喷水声", "acceptable_intermediate_concepts": ["滋滋"], "target_idiom_part": "孜孜", "gold_bridge_path": ["喷水", "滋滋", "孜孜"], "eviden... | |||
human_seed_0006__H4 | human_seed_0006 | human_seed_figure | images/human/孜孜不倦.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 泥古不化
B. 纹丝不动
C. 大可不必
D. 孜孜不倦
请只输出最可能的成语，不要输出选项字母，不要解释。 | 孜孜不倦 | [] | [
"泥古不化",
"纹丝不动",
"大可不必",
"孜孜不倦"
] | {"answer": "孜孜不倦", "visible_cues": [{"id": "P1", "content": "一个人很困"}, {"id": "P2", "content": "另一个人用水枪喷水"}, {"id": "P3", "content": "被喷水后变精神不困了"}], "bridge_units": [{"id": "B1", "image_concept": "喷水声", "acceptable_intermediate_concepts": ["滋滋"], "target_idiom_part": "孜孜", "gold_bridge_path": ["喷水", "滋滋", "孜孜"], "eviden... | |||
human_seed_0006__E0 | human_seed_0006 | human_seed_figure | images/human/孜孜不倦.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 孜孜不倦 | [] | [] | {"answer": "孜孜不倦", "visible_cues": [{"id": "P1", "content": "一个人很困"}, {"id": "P2", "content": "另一个人用水枪喷水"}, {"id": "P3", "content": "被喷水后变精神不困了"}], "bridge_units": [{"id": "B1", "image_concept": "喷水声", "acceptable_intermediate_concepts": ["滋滋"], "target_idiom_part": "孜孜", "gold_bridge_path": ["喷水", "滋滋", "孜孜"], "eviden... | |||
human_seed_0006__E1 | human_seed_0006 | human_seed_figure | images/human/孜孜不倦.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“孜孜不倦”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "孜孜不倦",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 孜孜不倦 | [] | [] | {"answer": "孜孜不倦", "visible_cues": [{"id": "P1", "content": "一个人很困"}, {"id": "P2", "content": "另一个人用水枪喷水"}, {"id": "P3", "content": "被喷水后变精神不困了"}], "bridge_units": [{"id": "B1", "image_concept": "喷水声", "acceptable_intermediate_concepts": ["滋滋"], "target_idiom_part": "孜孜", "gold_bridge_path": ["喷水", "滋滋", "孜孜"], "eviden... | |||
human_seed_0007__H0 | human_seed_0007 | human_seed_figure | images/human/背道而驰.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 背道而驰 | [] | [] | {"answer": "背道而驰", "visible_cues": [{"id": "P1", "content": "有人骑摩托飞奔"}, {"id": "P2", "content": "骑车的人背着道士"}, {"id": "P3", "content": "道士要求对方放他下来"}], "bridge_units": [{"id": "B1", "image_concept": "背着道士", "acceptable_intermediate_concepts": ["背道士"], "target_idiom_part": "背道", "gold_bridge_path": ["背着道士", "背道士", "背道"], "... | |||
human_seed_0007__H1 | human_seed_0007 | human_seed_figure | images/human/背道而驰.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 背道而驰 | [] | [] | {"answer": "背道而驰", "visible_cues": [{"id": "P1", "content": "有人骑摩托飞奔"}, {"id": "P2", "content": "骑车的人背着道士"}, {"id": "P3", "content": "道士要求对方放他下来"}], "bridge_units": [{"id": "B1", "image_concept": "背着道士", "acceptable_intermediate_concepts": ["背道士"], "target_idiom_part": "背道", "gold_bridge_path": ["背着道士", "背道士", "背道"], "... | |||
human_seed_0007__H4 | human_seed_0007 | human_seed_figure | images/human/背道而驰.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 头头是道
B. 背道而驰
C. 分道扬镳
D. 守株待兔
请只输出最可能的成语，不要输出选项字母，不要解释。 | 背道而驰 | [] | [
"头头是道",
"背道而驰",
"分道扬镳",
"守株待兔"
] | {"answer": "背道而驰", "visible_cues": [{"id": "P1", "content": "有人骑摩托飞奔"}, {"id": "P2", "content": "骑车的人背着道士"}, {"id": "P3", "content": "道士要求对方放他下来"}], "bridge_units": [{"id": "B1", "image_concept": "背着道士", "acceptable_intermediate_concepts": ["背道士"], "target_idiom_part": "背道", "gold_bridge_path": ["背着道士", "背道士", "背道"], "... | |||
human_seed_0007__E0 | human_seed_0007 | human_seed_figure | images/human/背道而驰.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 背道而驰 | [] | [] | {"answer": "背道而驰", "visible_cues": [{"id": "P1", "content": "有人骑摩托飞奔"}, {"id": "P2", "content": "骑车的人背着道士"}, {"id": "P3", "content": "道士要求对方放他下来"}], "bridge_units": [{"id": "B1", "image_concept": "背着道士", "acceptable_intermediate_concepts": ["背道士"], "target_idiom_part": "背道", "gold_bridge_path": ["背着道士", "背道士", "背道"], "... | |||
human_seed_0007__E1 | human_seed_0007 | human_seed_figure | images/human/背道而驰.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“背道而驰”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "背道而驰",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 背道而驰 | [] | [] | {"answer": "背道而驰", "visible_cues": [{"id": "P1", "content": "有人骑摩托飞奔"}, {"id": "P2", "content": "骑车的人背着道士"}, {"id": "P3", "content": "道士要求对方放他下来"}], "bridge_units": [{"id": "B1", "image_concept": "背着道士", "acceptable_intermediate_concepts": ["背道士"], "target_idiom_part": "背道", "gold_bridge_path": ["背着道士", "背道士", "背道"], "... | |||
human_seed_0008__H0 | human_seed_0008 | human_seed_figure | images/human/别开生面.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 别开生面 | [] | [] | {"answer": "别开生面", "visible_cues": [{"id": "P1", "content": "四个员工面前有几碗面"}, {"id": "P2", "content": "三碗是熟面，一碗是生面"}, {"id": "P3", "content": "蟹老板说员工被解雇/开除"}, {"id": "P4", "content": "海绵宝宝指着生面说不要解雇它"}], "bridge_units": [{"id": "B1", "image_concept": "不要开除", "acceptable_intermediate_concepts": ["别开除"], "target_idiom_part":... | |||
human_seed_0008__H1 | human_seed_0008 | human_seed_figure | images/human/别开生面.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 别开生面 | [] | [] | {"answer": "别开生面", "visible_cues": [{"id": "P1", "content": "四个员工面前有几碗面"}, {"id": "P2", "content": "三碗是熟面，一碗是生面"}, {"id": "P3", "content": "蟹老板说员工被解雇/开除"}, {"id": "P4", "content": "海绵宝宝指着生面说不要解雇它"}], "bridge_units": [{"id": "B1", "image_concept": "不要开除", "acceptable_intermediate_concepts": ["别开除"], "target_idiom_part":... | |||
human_seed_0008__H4 | human_seed_0008 | human_seed_figure | images/human/别开生面.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 茅塞顿开
B. 别开生面
C. 小鸟依人
D. 毛遂自荐
请只输出最可能的成语，不要输出选项字母，不要解释。 | 别开生面 | [] | [
"茅塞顿开",
"别开生面",
"小鸟依人",
"毛遂自荐"
] | {"answer": "别开生面", "visible_cues": [{"id": "P1", "content": "四个员工面前有几碗面"}, {"id": "P2", "content": "三碗是熟面，一碗是生面"}, {"id": "P3", "content": "蟹老板说员工被解雇/开除"}, {"id": "P4", "content": "海绵宝宝指着生面说不要解雇它"}], "bridge_units": [{"id": "B1", "image_concept": "不要开除", "acceptable_intermediate_concepts": ["别开除"], "target_idiom_part":... | |||
human_seed_0008__E0 | human_seed_0008 | human_seed_figure | images/human/别开生面.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 别开生面 | [] | [] | {"answer": "别开生面", "visible_cues": [{"id": "P1", "content": "四个员工面前有几碗面"}, {"id": "P2", "content": "三碗是熟面，一碗是生面"}, {"id": "P3", "content": "蟹老板说员工被解雇/开除"}, {"id": "P4", "content": "海绵宝宝指着生面说不要解雇它"}], "bridge_units": [{"id": "B1", "image_concept": "不要开除", "acceptable_intermediate_concepts": ["别开除"], "target_idiom_part":... | |||
human_seed_0008__E1 | human_seed_0008 | human_seed_figure | images/human/别开生面.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“别开生面”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "别开生面",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 别开生面 | [] | [] | {"answer": "别开生面", "visible_cues": [{"id": "P1", "content": "四个员工面前有几碗面"}, {"id": "P2", "content": "三碗是熟面，一碗是生面"}, {"id": "P3", "content": "蟹老板说员工被解雇/开除"}, {"id": "P4", "content": "海绵宝宝指着生面说不要解雇它"}], "bridge_units": [{"id": "B1", "image_concept": "不要开除", "acceptable_intermediate_concepts": ["别开除"], "target_idiom_part":... | |||
human_seed_0009__H0 | human_seed_0009 | human_seed_figure | images/human/勾心斗角.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 勾心斗角 | [] | [] | {"answer": "勾心斗角", "visible_cues": [{"id": "P1", "content": "一颗心出现在画面中"}, {"id": "P2", "content": "一个豆角在勾引或勾住心"}], "bridge_units": [{"id": "B1", "image_concept": "勾引一颗心", "acceptable_intermediate_concepts": [], "target_idiom_part": "勾心", "gold_bridge_path": ["勾引心", "勾心"], "evidence": ["P1", "P2"], "required": true}, {"... | |||
human_seed_0009__H1 | human_seed_0009 | human_seed_figure | images/human/勾心斗角.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 勾心斗角 | [] | [] | {"answer": "勾心斗角", "visible_cues": [{"id": "P1", "content": "一颗心出现在画面中"}, {"id": "P2", "content": "一个豆角在勾引或勾住心"}], "bridge_units": [{"id": "B1", "image_concept": "勾引一颗心", "acceptable_intermediate_concepts": [], "target_idiom_part": "勾心", "gold_bridge_path": ["勾引心", "勾心"], "evidence": ["P1", "P2"], "required": true}, {"... | |||
human_seed_0009__H4 | human_seed_0009 | human_seed_figure | images/human/勾心斗角.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 勾心斗角
B. 心服口服
C. 推心置腹
D. 口蜜腹剑
请只输出最可能的成语，不要输出选项字母，不要解释。 | 勾心斗角 | [] | [
"勾心斗角",
"心服口服",
"推心置腹",
"口蜜腹剑"
] | {"answer": "勾心斗角", "visible_cues": [{"id": "P1", "content": "一颗心出现在画面中"}, {"id": "P2", "content": "一个豆角在勾引或勾住心"}], "bridge_units": [{"id": "B1", "image_concept": "勾引一颗心", "acceptable_intermediate_concepts": [], "target_idiom_part": "勾心", "gold_bridge_path": ["勾引心", "勾心"], "evidence": ["P1", "P2"], "required": true}, {"... | |||
human_seed_0009__E0 | human_seed_0009 | human_seed_figure | images/human/勾心斗角.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 勾心斗角 | [] | [] | {"answer": "勾心斗角", "visible_cues": [{"id": "P1", "content": "一颗心出现在画面中"}, {"id": "P2", "content": "一个豆角在勾引或勾住心"}], "bridge_units": [{"id": "B1", "image_concept": "勾引一颗心", "acceptable_intermediate_concepts": [], "target_idiom_part": "勾心", "gold_bridge_path": ["勾引心", "勾心"], "evidence": ["P1", "P2"], "required": true}, {"... | |||
human_seed_0009__E1 | human_seed_0009 | human_seed_figure | images/human/勾心斗角.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“勾心斗角”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "勾心斗角",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 勾心斗角 | [] | [] | {"answer": "勾心斗角", "visible_cues": [{"id": "P1", "content": "一颗心出现在画面中"}, {"id": "P2", "content": "一个豆角在勾引或勾住心"}], "bridge_units": [{"id": "B1", "image_concept": "勾引一颗心", "acceptable_intermediate_concepts": [], "target_idiom_part": "勾心", "gold_bridge_path": ["勾引心", "勾心"], "evidence": ["P1", "P2"], "required": true}, {"... | |||
human_seed_0010__H0 | human_seed_0010 | human_seed_figure | images/human/积少成多.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 积少成多 | [] | [] | {"answer": "积少成多", "visible_cues": [{"id": "P1", "content": "一只鸡数量少"}, {"id": "P2", "content": "周围有很多橙子"}], "bridge_units": [{"id": "B1", "image_concept": "鸡少", "acceptable_intermediate_concepts": [], "target_idiom_part": "积少", "gold_bridge_path": ["鸡少", "积少"], "evidence": ["P1"], "required": true}, {"id": "B2", "image... | |||
human_seed_0010__H1 | human_seed_0010 | human_seed_figure | images/human/积少成多.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 积少成多 | [] | [] | {"answer": "积少成多", "visible_cues": [{"id": "P1", "content": "一只鸡数量少"}, {"id": "P2", "content": "周围有很多橙子"}], "bridge_units": [{"id": "B1", "image_concept": "鸡少", "acceptable_intermediate_concepts": [], "target_idiom_part": "积少", "gold_bridge_path": ["鸡少", "积少"], "evidence": ["P1"], "required": true}, {"id": "B2", "image... | |||
human_seed_0010__H4 | human_seed_0010 | human_seed_figure | images/human/积少成多.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 积少成多
B. 多管闲事
C. 毛遂自荐
D. 飞扬跋扈
请只输出最可能的成语，不要输出选项字母，不要解释。 | 积少成多 | [] | [
"积少成多",
"多管闲事",
"毛遂自荐",
"飞扬跋扈"
] | {"answer": "积少成多", "visible_cues": [{"id": "P1", "content": "一只鸡数量少"}, {"id": "P2", "content": "周围有很多橙子"}], "bridge_units": [{"id": "B1", "image_concept": "鸡少", "acceptable_intermediate_concepts": [], "target_idiom_part": "积少", "gold_bridge_path": ["鸡少", "积少"], "evidence": ["P1"], "required": true}, {"id": "B2", "image... | |||
human_seed_0010__E0 | human_seed_0010 | human_seed_figure | images/human/积少成多.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 积少成多 | [] | [] | {"answer": "积少成多", "visible_cues": [{"id": "P1", "content": "一只鸡数量少"}, {"id": "P2", "content": "周围有很多橙子"}], "bridge_units": [{"id": "B1", "image_concept": "鸡少", "acceptable_intermediate_concepts": [], "target_idiom_part": "积少", "gold_bridge_path": ["鸡少", "积少"], "evidence": ["P1"], "required": true}, {"id": "B2", "image... | |||
human_seed_0010__E1 | human_seed_0010 | human_seed_figure | images/human/积少成多.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“积少成多”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "积少成多",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 积少成多 | [] | [] | {"answer": "积少成多", "visible_cues": [{"id": "P1", "content": "一只鸡数量少"}, {"id": "P2", "content": "周围有很多橙子"}], "bridge_units": [{"id": "B1", "image_concept": "鸡少", "acceptable_intermediate_concepts": [], "target_idiom_part": "积少", "gold_bridge_path": ["鸡少", "积少"], "evidence": ["P1"], "required": true}, {"id": "B2", "image... | |||
human_seed_0011__H0 | human_seed_0011 | human_seed_figure | images/human/空前绝后.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 空前绝后 | [] | [] | {"answer": "空前绝后", "visible_cues": [{"id": "P1", "content": "图中出现大圣 / 孙悟空"}, {"id": "P2", "content": "有人提到练神功和葵花宝典"}, {"id": "P3", "content": "画面中有带血的刀"}, {"id": "P4", "content": "葵花宝典关联“欲练神功，必先自宫”"}, {"id": "P5", "content": "自宫后不能有后代"}], "bridge_units": [{"id": "B1", "image_concept": "孙悟空", "acceptable_intermediate_co... | |||
human_seed_0011__H1 | human_seed_0011 | human_seed_figure | images/human/空前绝后.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 空前绝后 | [] | [] | {"answer": "空前绝后", "visible_cues": [{"id": "P1", "content": "图中出现大圣 / 孙悟空"}, {"id": "P2", "content": "有人提到练神功和葵花宝典"}, {"id": "P3", "content": "画面中有带血的刀"}, {"id": "P4", "content": "葵花宝典关联“欲练神功，必先自宫”"}, {"id": "P5", "content": "自宫后不能有后代"}], "bridge_units": [{"id": "B1", "image_concept": "孙悟空", "acceptable_intermediate_co... | |||
human_seed_0011__H4 | human_seed_0011 | human_seed_figure | images/human/空前绝后.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 后患无穷
B. 死有余辜
C. 画饼充饥
D. 空前绝后
请只输出最可能的成语，不要输出选项字母，不要解释。 | 空前绝后 | [] | [
"后患无穷",
"死有余辜",
"画饼充饥",
"空前绝后"
] | {"answer": "空前绝后", "visible_cues": [{"id": "P1", "content": "图中出现大圣 / 孙悟空"}, {"id": "P2", "content": "有人提到练神功和葵花宝典"}, {"id": "P3", "content": "画面中有带血的刀"}, {"id": "P4", "content": "葵花宝典关联“欲练神功，必先自宫”"}, {"id": "P5", "content": "自宫后不能有后代"}], "bridge_units": [{"id": "B1", "image_concept": "孙悟空", "acceptable_intermediate_co... | |||
human_seed_0011__E0 | human_seed_0011 | human_seed_figure | images/human/空前绝后.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 空前绝后 | [] | [] | {"answer": "空前绝后", "visible_cues": [{"id": "P1", "content": "图中出现大圣 / 孙悟空"}, {"id": "P2", "content": "有人提到练神功和葵花宝典"}, {"id": "P3", "content": "画面中有带血的刀"}, {"id": "P4", "content": "葵花宝典关联“欲练神功，必先自宫”"}, {"id": "P5", "content": "自宫后不能有后代"}], "bridge_units": [{"id": "B1", "image_concept": "孙悟空", "acceptable_intermediate_co... | |||
human_seed_0011__E1 | human_seed_0011 | human_seed_figure | images/human/空前绝后.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“空前绝后”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "空前绝后",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 空前绝后 | [] | [] | {"answer": "空前绝后", "visible_cues": [{"id": "P1", "content": "图中出现大圣 / 孙悟空"}, {"id": "P2", "content": "有人提到练神功和葵花宝典"}, {"id": "P3", "content": "画面中有带血的刀"}, {"id": "P4", "content": "葵花宝典关联“欲练神功，必先自宫”"}, {"id": "P5", "content": "自宫后不能有后代"}], "bridge_units": [{"id": "B1", "image_concept": "孙悟空", "acceptable_intermediate_co... | |||
human_seed_0012__H0 | human_seed_0012 | human_seed_figure | images/human/倾家荡产.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 倾家荡产 | [] | [] | {"answer": "倾家荡产", "visible_cues": [{"id": "P1", "content": "一个人称另一个人为亲家"}, {"id": "P2", "content": "另一个人在荡秋千"}, {"id": "P3", "content": "秋千底部有铲子"}], "bridge_units": [{"id": "B1", "image_concept": "亲家", "acceptable_intermediate_concepts": [], "target_idiom_part": "倾家", "gold_bridge_path": ["亲家", "倾家"], "evidence": ["P1... | |||
human_seed_0012__H1 | human_seed_0012 | human_seed_figure | images/human/倾家荡产.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 倾家荡产 | [] | [] | {"answer": "倾家荡产", "visible_cues": [{"id": "P1", "content": "一个人称另一个人为亲家"}, {"id": "P2", "content": "另一个人在荡秋千"}, {"id": "P3", "content": "秋千底部有铲子"}], "bridge_units": [{"id": "B1", "image_concept": "亲家", "acceptable_intermediate_concepts": [], "target_idiom_part": "倾家", "gold_bridge_path": ["亲家", "倾家"], "evidence": ["P1... | |||
human_seed_0012__H4 | human_seed_0012 | human_seed_figure | images/human/倾家荡产.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 倾家荡产
B. 半路出家
C. 亡羊补牢
D. 素不相识
请只输出最可能的成语，不要输出选项字母，不要解释。 | 倾家荡产 | [] | [
"倾家荡产",
"半路出家",
"亡羊补牢",
"素不相识"
] | {"answer": "倾家荡产", "visible_cues": [{"id": "P1", "content": "一个人称另一个人为亲家"}, {"id": "P2", "content": "另一个人在荡秋千"}, {"id": "P3", "content": "秋千底部有铲子"}], "bridge_units": [{"id": "B1", "image_concept": "亲家", "acceptable_intermediate_concepts": [], "target_idiom_part": "倾家", "gold_bridge_path": ["亲家", "倾家"], "evidence": ["P1... | |||
human_seed_0012__E0 | human_seed_0012 | human_seed_figure | images/human/倾家荡产.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 倾家荡产 | [] | [] | {"answer": "倾家荡产", "visible_cues": [{"id": "P1", "content": "一个人称另一个人为亲家"}, {"id": "P2", "content": "另一个人在荡秋千"}, {"id": "P3", "content": "秋千底部有铲子"}], "bridge_units": [{"id": "B1", "image_concept": "亲家", "acceptable_intermediate_concepts": [], "target_idiom_part": "倾家", "gold_bridge_path": ["亲家", "倾家"], "evidence": ["P1... | |||
human_seed_0012__E1 | human_seed_0012 | human_seed_figure | images/human/倾家荡产.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“倾家荡产”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "倾家荡产",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 倾家荡产 | [] | [] | {"answer": "倾家荡产", "visible_cues": [{"id": "P1", "content": "一个人称另一个人为亲家"}, {"id": "P2", "content": "另一个人在荡秋千"}, {"id": "P3", "content": "秋千底部有铲子"}], "bridge_units": [{"id": "B1", "image_concept": "亲家", "acceptable_intermediate_concepts": [], "target_idiom_part": "倾家", "gold_bridge_path": ["亲家", "倾家"], "evidence": ["P1... | |||
human_seed_0013__H0 | human_seed_0013 | human_seed_figure | images/human/死有余辜.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 死有余辜 | [] | [] | {"answer": "死有余辜", "visible_cues": [{"id": "P1", "content": "一坨屎"}, {"id": "P2", "content": "屎里面有鱼"}, {"id": "P3", "content": "屎里面有蘑菇"}], "bridge_units": [{"id": "B1", "image_concept": "屎", "acceptable_intermediate_concepts": [], "target_idiom_part": "死", "gold_bridge_path": ["屎", "死"], "evidence": ["P1"], "required": ... | |||
human_seed_0013__H1 | human_seed_0013 | human_seed_figure | images/human/死有余辜.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 死有余辜 | [] | [] | {"answer": "死有余辜", "visible_cues": [{"id": "P1", "content": "一坨屎"}, {"id": "P2", "content": "屎里面有鱼"}, {"id": "P3", "content": "屎里面有蘑菇"}], "bridge_units": [{"id": "B1", "image_concept": "屎", "acceptable_intermediate_concepts": [], "target_idiom_part": "死", "gold_bridge_path": ["屎", "死"], "evidence": ["P1"], "required": ... | |||
human_seed_0013__H4 | human_seed_0013 | human_seed_figure | images/human/死有余辜.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 各有千秋
B. 有勇无谋
C. 死有余辜
D. 路人皆知
请只输出最可能的成语，不要输出选项字母，不要解释。 | 死有余辜 | [] | [
"各有千秋",
"有勇无谋",
"死有余辜",
"路人皆知"
] | {"answer": "死有余辜", "visible_cues": [{"id": "P1", "content": "一坨屎"}, {"id": "P2", "content": "屎里面有鱼"}, {"id": "P3", "content": "屎里面有蘑菇"}], "bridge_units": [{"id": "B1", "image_concept": "屎", "acceptable_intermediate_concepts": [], "target_idiom_part": "死", "gold_bridge_path": ["屎", "死"], "evidence": ["P1"], "required": ... | |||
human_seed_0013__E0 | human_seed_0013 | human_seed_figure | images/human/死有余辜.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 死有余辜 | [] | [] | {"answer": "死有余辜", "visible_cues": [{"id": "P1", "content": "一坨屎"}, {"id": "P2", "content": "屎里面有鱼"}, {"id": "P3", "content": "屎里面有蘑菇"}], "bridge_units": [{"id": "B1", "image_concept": "屎", "acceptable_intermediate_concepts": [], "target_idiom_part": "死", "gold_bridge_path": ["屎", "死"], "evidence": ["P1"], "required": ... | |||
human_seed_0013__E1 | human_seed_0013 | human_seed_figure | images/human/死有余辜.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“死有余辜”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "死有余辜",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 死有余辜 | [] | [] | {"answer": "死有余辜", "visible_cues": [{"id": "P1", "content": "一坨屎"}, {"id": "P2", "content": "屎里面有鱼"}, {"id": "P3", "content": "屎里面有蘑菇"}], "bridge_units": [{"id": "B1", "image_concept": "屎", "acceptable_intermediate_concepts": [], "target_idiom_part": "死", "gold_bridge_path": ["屎", "死"], "evidence": ["P1"], "required": ... | |||
human_seed_0014__H0 | human_seed_0014 | human_seed_figure | images/human/有勇无谋.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 有勇无谋 | [] | [] | {"answer": "有勇无谋", "visible_cues": [{"id": "P1", "content": "有人在水里游泳"}, {"id": "P2", "content": "岸上的人有眼眸"}, {"id": "P3", "content": "水里游泳的人没有眼眸"}], "bridge_units": [{"id": "B1", "image_concept": "游泳", "acceptable_intermediate_concepts": [], "target_idiom_part": "有勇", "gold_bridge_path": ["游泳", "有勇"], "evidence": ["P1"]... | |||
human_seed_0014__H1 | human_seed_0014 | human_seed_figure | images/human/有勇无谋.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 有勇无谋 | [] | [] | {"answer": "有勇无谋", "visible_cues": [{"id": "P1", "content": "有人在水里游泳"}, {"id": "P2", "content": "岸上的人有眼眸"}, {"id": "P3", "content": "水里游泳的人没有眼眸"}], "bridge_units": [{"id": "B1", "image_concept": "游泳", "acceptable_intermediate_concepts": [], "target_idiom_part": "有勇", "gold_bridge_path": ["游泳", "有勇"], "evidence": ["P1"]... | |||
human_seed_0014__H4 | human_seed_0014 | human_seed_figure | images/human/有勇无谋.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 有勇无谋
B. 各有千秋
C. 后患无穷
D. 死有余辜
请只输出最可能的成语，不要输出选项字母，不要解释。 | 有勇无谋 | [] | [
"有勇无谋",
"各有千秋",
"后患无穷",
"死有余辜"
] | {"answer": "有勇无谋", "visible_cues": [{"id": "P1", "content": "有人在水里游泳"}, {"id": "P2", "content": "岸上的人有眼眸"}, {"id": "P3", "content": "水里游泳的人没有眼眸"}], "bridge_units": [{"id": "B1", "image_concept": "游泳", "acceptable_intermediate_concepts": [], "target_idiom_part": "有勇", "gold_bridge_path": ["游泳", "有勇"], "evidence": ["P1"]... | |||
human_seed_0014__E0 | human_seed_0014 | human_seed_figure | images/human/有勇无谋.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 有勇无谋 | [] | [] | {"answer": "有勇无谋", "visible_cues": [{"id": "P1", "content": "有人在水里游泳"}, {"id": "P2", "content": "岸上的人有眼眸"}, {"id": "P3", "content": "水里游泳的人没有眼眸"}], "bridge_units": [{"id": "B1", "image_concept": "游泳", "acceptable_intermediate_concepts": [], "target_idiom_part": "有勇", "gold_bridge_path": ["游泳", "有勇"], "evidence": ["P1"]... | |||
human_seed_0014__E1 | human_seed_0014 | human_seed_figure | images/human/有勇无谋.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“有勇无谋”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "有勇无谋",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 有勇无谋 | [] | [] | {"answer": "有勇无谋", "visible_cues": [{"id": "P1", "content": "有人在水里游泳"}, {"id": "P2", "content": "岸上的人有眼眸"}, {"id": "P3", "content": "水里游泳的人没有眼眸"}], "bridge_units": [{"id": "B1", "image_concept": "游泳", "acceptable_intermediate_concepts": [], "target_idiom_part": "有勇", "gold_bridge_path": ["游泳", "有勇"], "evidence": ["P1"]... | |||
human_seed_0015__H0 | human_seed_0015 | human_seed_figure | images/human/愚昧无知.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 愚昧无知 | [] | [] | {"answer": "愚昧无知", "visible_cues": [{"id": "P1", "content": "四条鱼组成家庭场景"}, {"id": "P2", "content": "小的雌性鱼像妹妹"}, {"id": "P3", "content": "其他鱼都有芝士而她没有"}], "bridge_units": [{"id": "B1", "image_concept": "鱼妹妹", "acceptable_intermediate_concepts": ["鱼妹"], "target_idiom_part": "愚昧", "gold_bridge_path": ["鱼妹", "愚昧"], "evidence... | |||
human_seed_0015__H1 | human_seed_0015 | human_seed_figure | images/human/愚昧无知.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 愚昧无知 | [] | [] | {"answer": "愚昧无知", "visible_cues": [{"id": "P1", "content": "四条鱼组成家庭场景"}, {"id": "P2", "content": "小的雌性鱼像妹妹"}, {"id": "P3", "content": "其他鱼都有芝士而她没有"}], "bridge_units": [{"id": "B1", "image_concept": "鱼妹妹", "acceptable_intermediate_concepts": ["鱼妹"], "target_idiom_part": "愚昧", "gold_bridge_path": ["鱼妹", "愚昧"], "evidence... | |||
human_seed_0015__H4 | human_seed_0015 | human_seed_figure | images/human/愚昧无知.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 有勇无谋
B. 后患无穷
C. 愚昧无知
D. 路人皆知
请只输出最可能的成语，不要输出选项字母，不要解释。 | 愚昧无知 | [] | [
"有勇无谋",
"后患无穷",
"愚昧无知",
"路人皆知"
] | {"answer": "愚昧无知", "visible_cues": [{"id": "P1", "content": "四条鱼组成家庭场景"}, {"id": "P2", "content": "小的雌性鱼像妹妹"}, {"id": "P3", "content": "其他鱼都有芝士而她没有"}], "bridge_units": [{"id": "B1", "image_concept": "鱼妹妹", "acceptable_intermediate_concepts": ["鱼妹"], "target_idiom_part": "愚昧", "gold_bridge_path": ["鱼妹", "愚昧"], "evidence... | |||
human_seed_0015__E0 | human_seed_0015 | human_seed_figure | images/human/愚昧无知.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 愚昧无知 | [] | [] | {"answer": "愚昧无知", "visible_cues": [{"id": "P1", "content": "四条鱼组成家庭场景"}, {"id": "P2", "content": "小的雌性鱼像妹妹"}, {"id": "P3", "content": "其他鱼都有芝士而她没有"}], "bridge_units": [{"id": "B1", "image_concept": "鱼妹妹", "acceptable_intermediate_concepts": ["鱼妹"], "target_idiom_part": "愚昧", "gold_bridge_path": ["鱼妹", "愚昧"], "evidence... | |||
human_seed_0015__E1 | human_seed_0015 | human_seed_figure | images/human/愚昧无知.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“愚昧无知”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "愚昧无知",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 愚昧无知 | [] | [] | {"answer": "愚昧无知", "visible_cues": [{"id": "P1", "content": "四条鱼组成家庭场景"}, {"id": "P2", "content": "小的雌性鱼像妹妹"}, {"id": "P3", "content": "其他鱼都有芝士而她没有"}], "bridge_units": [{"id": "B1", "image_concept": "鱼妹妹", "acceptable_intermediate_concepts": ["鱼妹"], "target_idiom_part": "愚昧", "gold_bridge_path": ["鱼妹", "愚昧"], "evidence... | |||
human_seed_0016__H0 | human_seed_0016 | human_seed_figure | images/human/正人君子.png | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 正人君子 | [] | [] | {"answer": "正人君子", "visible_cues": [{"id": "P1", "content": "法庭上出现证人"}, {"id": "P2", "content": "证人是蘑菇形象"}, {"id": "P3", "content": "蘑菇也可称菌子"}], "bridge_units": [{"id": "B1", "image_concept": "证人", "acceptable_intermediate_concepts": [], "target_idiom_part": "正人", "gold_bridge_path": ["证人", "正人"], "evidence": ["P1"], "... | |||
human_seed_0016__H1 | human_seed_0016 | human_seed_figure | images/human/正人君子.png | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 正人君子 | [] | [] | {"answer": "正人君子", "visible_cues": [{"id": "P1", "content": "法庭上出现证人"}, {"id": "P2", "content": "证人是蘑菇形象"}, {"id": "P3", "content": "蘑菇也可称菌子"}], "bridge_units": [{"id": "B1", "image_concept": "证人", "acceptable_intermediate_concepts": [], "target_idiom_part": "正人", "gold_bridge_path": ["证人", "正人"], "evidence": ["P1"], "... | |||
human_seed_0016__H4 | human_seed_0016 | human_seed_figure | images/human/正人君子.png | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 平易近人
B. 路人皆知
C. 小鸟依人
D. 正人君子
请只输出最可能的成语，不要输出选项字母，不要解释。 | 正人君子 | [] | [
"平易近人",
"路人皆知",
"小鸟依人",
"正人君子"
] | {"answer": "正人君子", "visible_cues": [{"id": "P1", "content": "法庭上出现证人"}, {"id": "P2", "content": "证人是蘑菇形象"}, {"id": "P3", "content": "蘑菇也可称菌子"}], "bridge_units": [{"id": "B1", "image_concept": "证人", "acceptable_intermediate_concepts": [], "target_idiom_part": "正人", "gold_bridge_path": ["证人", "正人"], "evidence": ["P1"], "... | |||
human_seed_0016__E0 | human_seed_0016 | human_seed_figure | images/human/正人君子.png | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 正人君子 | [] | [] | {"answer": "正人君子", "visible_cues": [{"id": "P1", "content": "法庭上出现证人"}, {"id": "P2", "content": "证人是蘑菇形象"}, {"id": "P3", "content": "蘑菇也可称菌子"}], "bridge_units": [{"id": "B1", "image_concept": "证人", "acceptable_intermediate_concepts": [], "target_idiom_part": "正人", "gold_bridge_path": ["证人", "正人"], "evidence": ["P1"], "... | |||
human_seed_0016__E1 | human_seed_0016 | human_seed_figure | images/human/正人君子.png | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“正人君子”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "正人君子",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 正人君子 | [] | [] | {"answer": "正人君子", "visible_cues": [{"id": "P1", "content": "法庭上出现证人"}, {"id": "P2", "content": "证人是蘑菇形象"}, {"id": "P3", "content": "蘑菇也可称菌子"}], "bridge_units": [{"id": "B1", "image_concept": "证人", "acceptable_intermediate_concepts": [], "target_idiom_part": "正人", "gold_bridge_path": ["证人", "正人"], "evidence": ["P1"], "... | |||
human_seed_0017__H0 | human_seed_0017 | human_seed_figure | images/human/半路出家.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 半路出家 | [] | [] | {"answer": "半路出家", "visible_cues": [{"id": "P1", "content": "图中出现路飞"}, {"id": "P2", "content": "路飞一半保持原样"}, {"id": "P3", "content": "另一半变成和尚形象"}], "bridge_units": [{"id": "B1", "image_concept": "半个路飞", "acceptable_intermediate_concepts": ["半路"], "target_idiom_part": "半路", "gold_bridge_path": ["半个路飞", "半路"], "evidence":... | |||
human_seed_0017__H1 | human_seed_0017 | human_seed_figure | images/human/半路出家.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 半路出家 | [] | [] | {"answer": "半路出家", "visible_cues": [{"id": "P1", "content": "图中出现路飞"}, {"id": "P2", "content": "路飞一半保持原样"}, {"id": "P3", "content": "另一半变成和尚形象"}], "bridge_units": [{"id": "B1", "image_concept": "半个路飞", "acceptable_intermediate_concepts": ["半路"], "target_idiom_part": "半路", "gold_bridge_path": ["半个路飞", "半路"], "evidence":... | |||
human_seed_0017__H4 | human_seed_0017 | human_seed_figure | images/human/半路出家.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 路人皆知
B. 倾家荡产
C. 程门立雪
D. 半路出家
请只输出最可能的成语，不要输出选项字母，不要解释。 | 半路出家 | [] | [
"路人皆知",
"倾家荡产",
"程门立雪",
"半路出家"
] | {"answer": "半路出家", "visible_cues": [{"id": "P1", "content": "图中出现路飞"}, {"id": "P2", "content": "路飞一半保持原样"}, {"id": "P3", "content": "另一半变成和尚形象"}], "bridge_units": [{"id": "B1", "image_concept": "半个路飞", "acceptable_intermediate_concepts": ["半路"], "target_idiom_part": "半路", "gold_bridge_path": ["半个路飞", "半路"], "evidence":... | |||
human_seed_0017__E0 | human_seed_0017 | human_seed_figure | images/human/半路出家.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 半路出家 | [] | [] | {"answer": "半路出家", "visible_cues": [{"id": "P1", "content": "图中出现路飞"}, {"id": "P2", "content": "路飞一半保持原样"}, {"id": "P3", "content": "另一半变成和尚形象"}], "bridge_units": [{"id": "B1", "image_concept": "半个路飞", "acceptable_intermediate_concepts": ["半路"], "target_idiom_part": "半路", "gold_bridge_path": ["半个路飞", "半路"], "evidence":... | |||
human_seed_0017__E1 | human_seed_0017 | human_seed_figure | images/human/半路出家.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“半路出家”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "半路出家",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 半路出家 | [] | [] | {"answer": "半路出家", "visible_cues": [{"id": "P1", "content": "图中出现路飞"}, {"id": "P2", "content": "路飞一半保持原样"}, {"id": "P3", "content": "另一半变成和尚形象"}], "bridge_units": [{"id": "B1", "image_concept": "半个路飞", "acceptable_intermediate_concepts": ["半路"], "target_idiom_part": "半路", "gold_bridge_path": ["半个路飞", "半路"], "evidence":... | |||
human_seed_0018__H0 | human_seed_0018 | human_seed_figure | images/human/大可不必.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 大可不必 | [] | [] | {"answer": "大可不必", "visible_cues": [{"id": "P1", "content": "有人在卖可乐"}, {"id": "P2", "content": "顾客拒绝大杯可乐并说小杯即可"}], "bridge_units": [{"id": "B1", "image_concept": "大杯可乐", "acceptable_intermediate_concepts": ["大可"], "target_idiom_part": "大可", "gold_bridge_path": ["大杯可乐", "大可"], "evidence": ["P1", "P2"], "required": true}... | |||
human_seed_0018__H1 | human_seed_0018 | human_seed_figure | images/human/大可不必.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 大可不必 | [] | [] | {"answer": "大可不必", "visible_cues": [{"id": "P1", "content": "有人在卖可乐"}, {"id": "P2", "content": "顾客拒绝大杯可乐并说小杯即可"}], "bridge_units": [{"id": "B1", "image_concept": "大杯可乐", "acceptable_intermediate_concepts": ["大可"], "target_idiom_part": "大可", "gold_bridge_path": ["大杯可乐", "大可"], "evidence": ["P1", "P2"], "required": true}... | |||
human_seed_0018__H4 | human_seed_0018 | human_seed_figure | images/human/大可不必.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 怒不可遏
B. 大可不必
C. 素不相识
D. 孜孜不倦
请只输出最可能的成语，不要输出选项字母，不要解释。 | 大可不必 | [] | [
"怒不可遏",
"大可不必",
"素不相识",
"孜孜不倦"
] | {"answer": "大可不必", "visible_cues": [{"id": "P1", "content": "有人在卖可乐"}, {"id": "P2", "content": "顾客拒绝大杯可乐并说小杯即可"}], "bridge_units": [{"id": "B1", "image_concept": "大杯可乐", "acceptable_intermediate_concepts": ["大可"], "target_idiom_part": "大可", "gold_bridge_path": ["大杯可乐", "大可"], "evidence": ["P1", "P2"], "required": true}... | |||
human_seed_0018__E0 | human_seed_0018 | human_seed_figure | images/human/大可不必.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 大可不必 | [] | [] | {"answer": "大可不必", "visible_cues": [{"id": "P1", "content": "有人在卖可乐"}, {"id": "P2", "content": "顾客拒绝大杯可乐并说小杯即可"}], "bridge_units": [{"id": "B1", "image_concept": "大杯可乐", "acceptable_intermediate_concepts": ["大可"], "target_idiom_part": "大可", "gold_bridge_path": ["大杯可乐", "大可"], "evidence": ["P1", "P2"], "required": true}... | |||
human_seed_0018__E1 | human_seed_0018 | human_seed_figure | images/human/大可不必.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“大可不必”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "大可不必",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 大可不必 | [] | [] | {"answer": "大可不必", "visible_cues": [{"id": "P1", "content": "有人在卖可乐"}, {"id": "P2", "content": "顾客拒绝大杯可乐并说小杯即可"}], "bridge_units": [{"id": "B1", "image_concept": "大杯可乐", "acceptable_intermediate_concepts": ["大可"], "target_idiom_part": "大可", "gold_bridge_path": ["大杯可乐", "大可"], "evidence": ["P1", "P2"], "required": true}... | |||
human_seed_0019__H0 | human_seed_0019 | human_seed_figure | images/human/单刀赴会.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 单刀赴会 | [] | [] | {"answer": "单刀赴会", "visible_cues": [{"id": "P1", "content": "两个人准备约会"}, {"id": "P2", "content": "男方只带了一美元"}, {"id": "P3", "content": "美元可称 dollar / 美刀"}], "bridge_units": [{"id": "B1", "image_concept": "一美元", "acceptable_intermediate_concepts": ["一刀", "美刀"], "target_idiom_part": "单刀", "gold_bridge_path": ["一美元", "一刀", ... | |||
human_seed_0019__H1 | human_seed_0019 | human_seed_figure | images/human/单刀赴会.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 单刀赴会 | [] | [] | {"answer": "单刀赴会", "visible_cues": [{"id": "P1", "content": "两个人准备约会"}, {"id": "P2", "content": "男方只带了一美元"}, {"id": "P3", "content": "美元可称 dollar / 美刀"}], "bridge_units": [{"id": "B1", "image_concept": "一美元", "acceptable_intermediate_concepts": ["一刀", "美刀"], "target_idiom_part": "单刀", "gold_bridge_path": ["一美元", "一刀", ... | |||
human_seed_0019__H4 | human_seed_0019 | human_seed_figure | images/human/单刀赴会.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 叹为观止
B. 单刀赴会
C. 一网打尽
D. 趋炎附势
请只输出最可能的成语，不要输出选项字母，不要解释。 | 单刀赴会 | [] | [
"叹为观止",
"单刀赴会",
"一网打尽",
"趋炎附势"
] | {"answer": "单刀赴会", "visible_cues": [{"id": "P1", "content": "两个人准备约会"}, {"id": "P2", "content": "男方只带了一美元"}, {"id": "P3", "content": "美元可称 dollar / 美刀"}], "bridge_units": [{"id": "B1", "image_concept": "一美元", "acceptable_intermediate_concepts": ["一刀", "美刀"], "target_idiom_part": "单刀", "gold_bridge_path": ["一美元", "一刀", ... | |||
human_seed_0019__E0 | human_seed_0019 | human_seed_figure | images/human/单刀赴会.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 单刀赴会 | [] | [] | {"answer": "单刀赴会", "visible_cues": [{"id": "P1", "content": "两个人准备约会"}, {"id": "P2", "content": "男方只带了一美元"}, {"id": "P3", "content": "美元可称 dollar / 美刀"}], "bridge_units": [{"id": "B1", "image_concept": "一美元", "acceptable_intermediate_concepts": ["一刀", "美刀"], "target_idiom_part": "单刀", "gold_bridge_path": ["一美元", "一刀", ... | |||
human_seed_0019__E1 | human_seed_0019 | human_seed_figure | images/human/单刀赴会.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“单刀赴会”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "单刀赴会",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 单刀赴会 | [] | [] | {"answer": "单刀赴会", "visible_cues": [{"id": "P1", "content": "两个人准备约会"}, {"id": "P2", "content": "男方只带了一美元"}, {"id": "P3", "content": "美元可称 dollar / 美刀"}], "bridge_units": [{"id": "B1", "image_concept": "一美元", "acceptable_intermediate_concepts": ["一刀", "美刀"], "target_idiom_part": "单刀", "gold_bridge_path": ["一美元", "一刀", ... | |||
human_seed_0020__H0 | human_seed_0020 | human_seed_figure | images/human/断章取义.jpg | H0 | image_only_idiom_answer | 你会看到一张图。请猜它对应的中文成语。只输出一个成语，不要解释。 | 断章取义 | [] | [] | {"answer": "断章取义", "visible_cues": [{"id": "P1", "content": "一只章鱼断了触手"}, {"id": "P2", "content": "章鱼去银行取钱"}, {"id": "P3", "content": "取款金额是一个亿"}], "bridge_units": [{"id": "B1", "image_concept": "断了触手的章鱼", "acceptable_intermediate_concepts": ["断章鱼"], "target_idiom_part": "断章", "gold_bridge_path": ["断了的章鱼", "断章鱼", "断章"],... | |||
human_seed_0020__H1 | human_seed_0020 | human_seed_figure | images/human/断章取义.jpg | H1 | cross_concept_hint_idiom_answer | 你会看到一张图。这张图可能通过谐音、拆字、替换、角色联想、物品联想等跨概念方式表达一个中文成语。请根据图像线索猜成语。只输出一个成语，不要解释。 | 断章取义 | [] | [] | {"answer": "断章取义", "visible_cues": [{"id": "P1", "content": "一只章鱼断了触手"}, {"id": "P2", "content": "章鱼去银行取钱"}, {"id": "P3", "content": "取款金额是一个亿"}], "bridge_units": [{"id": "B1", "image_concept": "断了触手的章鱼", "acceptable_intermediate_concepts": ["断章鱼"], "target_idiom_part": "断章", "gold_bridge_path": ["断了的章鱼", "断章鱼", "断章"],... | |||
human_seed_0020__H4 | human_seed_0020 | human_seed_figure | images/human/断章取义.jpg | H4 | candidate_constrained_idiom_answer | 你会看到一张图。它通过跨概念方式表达下面候选成语之一。
A. 以德服人
B. 口蜜腹剑
C. 断章取义
D. 半路出家
请只输出最可能的成语，不要输出选项字母，不要解释。 | 断章取义 | [] | [
"以德服人",
"口蜜腹剑",
"断章取义",
"半路出家"
] | {"answer": "断章取义", "visible_cues": [{"id": "P1", "content": "一只章鱼断了触手"}, {"id": "P2", "content": "章鱼去银行取钱"}, {"id": "P3", "content": "取款金额是一个亿"}], "bridge_units": [{"id": "B1", "image_concept": "断了触手的章鱼", "acceptable_intermediate_concepts": ["断章鱼"], "target_idiom_part": "断章", "gold_bridge_path": ["断了的章鱼", "断章鱼", "断章"],... | |||
human_seed_0020__E0 | human_seed_0020 | human_seed_figure | images/human/断章取义.jpg | E0 | free_answer_with_explanation | 你会看到一张图。这张图是某个中文成语的跨概念创意表达。
请先给出你认为的成语，再解释你如何从图像线索推到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "成语",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 断章取义 | [] | [] | {"answer": "断章取义", "visible_cues": [{"id": "P1", "content": "一只章鱼断了触手"}, {"id": "P2", "content": "章鱼去银行取钱"}, {"id": "P3", "content": "取款金额是一个亿"}], "bridge_units": [{"id": "B1", "image_concept": "断了触手的章鱼", "acceptable_intermediate_concepts": ["断章鱼"], "target_idiom_part": "断章", "gold_bridge_path": ["断了的章鱼", "断章鱼", "断章"],... | |||
human_seed_0020__E1 | human_seed_0020 | human_seed_figure | images/human/断章取义.jpg | E1 | gold_answer_explanation | 你会看到一张图。已知它表达的成语是“断章取义”。
请解释图像中的关键线索如何通过跨概念桥接到这个成语。
必须返回 JSON，不要输出 JSON 以外的文字，格式如下：
{
"answer": "断章取义",
"perceptual_cues": ["你看见的关键图像线索"],
"bridge_analysis": [
{"visual_or_textual_cue": "线索", "bridged_concept": "被联想到的概念", "idiom_part": "对应的成语片段"}
],
"final_rationale": "一句话说明整体推理"
} | 断章取义 | [] | [] | {"answer": "断章取义", "visible_cues": [{"id": "P1", "content": "一只章鱼断了触手"}, {"id": "P2", "content": "章鱼去银行取钱"}, {"id": "P3", "content": "取款金额是一个亿"}], "bridge_units": [{"id": "B1", "image_concept": "断了触手的章鱼", "acceptable_intermediate_concepts": ["断章鱼"], "target_idiom_part": "断章", "gold_bridge_path": ["断了的章鱼", "断章鱼", "断章"],... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.