Files
nagi-bench/models/gpt-5-5-pro.json
T
Jesse ZhangandClaude Opus 4.8 16f528892c Tame GPT-5.5 Pro skeleton-watch blinding bloom glow
UnrealBloomPass used threshold 0.18 / strength 0.72, so with ACES tone
mapping and several strong lights nearly every highlight and lume surface
bloomed into one over-exposed white blob that washed out the whole dial.
Tame it to strength 0.28 / radius 0.45 / threshold 0.9 so the glow only
haloes the lume markers and the watch is legible again. Documented in the
run note (now a 2-line fix).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-19 01:01:18 +01:00

32 lines
4.4 KiB
JSON
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
{
"label": "GPT-5.5 Pro",
"vendor": "OpenAI",
"harness": "ChatGPT Web",
"effort": "Extended Pro",
"artifactDir": "gpt-5-5-pro/chatgpt-web-extended-pro",
"order": 20,
"runs": {
"mythos-craft": {
"note": {
"zh": "以 GPT-5.5 Pro(开启 Extended Pro)生成;本页为修复版(仅改 1 行)。问题:14 张贴图排成 8×2 图集、第二行只画了 6 格,UV 按左上原点计算却未设 flipY=false,两行贴图上下互换采样。修复前:草地顶面显示成黑曜石(地面全黑),树叶、沙子采到未绘制的空格(大片纯黑),放置方块的贴图互相串行、颜色诡异似反色。原因:CanvasTexture 默认 flipY=true,与自上而下的图集布局不匹配。加上 flipY=false 后地形与方块颜色全部正常。",
"en": "Generated with GPT-5.5 Pro (Extended Pro enabled); fixed build (1 line changed). Bug: 14 tiles in an 8x2 atlas with only 6 slots drawn in row two, UVs computed from a top-left origin without flipY=false, so the two rows swapped on sampling. Before the fix: grass tops rendered as obsidian (black ground), leaves and sand hit undrawn slots (large black patches), and placed blocks wore each other's textures with inverted-looking colors. Cause: CanvasTexture defaults to flipY=true, mismatching the top-down atlas layout. With flipY=false everything renders correctly."
},
"contributor": "Nagi-ovo"
},
"pelican-cycling": {
"note": {
"zh": "在 ChatGPT 网页版以 GPT-5.5 Pro(开启 Extended Pro)一次生成。",
"en": "Generated in one shot in the ChatGPT web app on GPT-5.5 Pro (Extended Pro enabled)."
},
"contributor": "Nagi-ovo"
},
"skeleton-watch": {
"note": {
"zh": "在 ChatGPT 网页版以 GPT-5.5 Pro(开启 Extended Pro)一次生成;本页为修复版(改 2 行)。问题一:发条与游丝用 createSpiralTube 构建(760 行),管线段数写成 Math.max(96, turns * 38),turns 为 5.8 / 6.4 时算出 220.4 / 243.2 的小数段数。修复前:TubeGeometry 末尾会以该段数作下标取 Frenet 帧 frames.normals[220.4],得到 undefined 后读取 .x 抛 TypeError,初始化中断,页面卡在「Assembling jeweled movement...」加载遮罩永不消失。原因:模型把管线段数当连续值传入,未取整。改为 Math.round(turns * 38) 后段数为整数,场景正常加载、遮罩按时隐藏。问题二:场景中心有一团异常刺眼的辉光,几乎把整块表盘糊成纯白,什么都看不清。原因:UnrealBloomPass(340 行)阈值过低(threshold 0.18)且强度偏高(strength 0.72),叠加 ACES 色调映射与多盏高强度灯光,几乎所有金属高光与夜光都被纳入泛光并扩散成一大团过曝白光。为保护测试者的眼睛,把 bloom 参数收敛为 strength 0.28 / radius 0.45 / threshold 0.9:泛光只保留夜光时标的柔和边缘,机芯、刻度、指针重新清晰可读。",
"en": "Generated in one shot in the ChatGPT web app on GPT-5.5 Pro (Extended Pro enabled); fixed build (2 lines changed). Bug 1: the mainspring and hairspring are built via createSpiralTube (line 760) with the tube segment count written as Math.max(96, turns * 38), so turns of 5.8 / 6.4 yield fractional counts 220.4 / 243.2. Before the fix: TubeGeometry's trailing generateSegment indexes the Frenet-frame arrays with that count (frames.normals[220.4]), gets undefined and reads .x, throwing a TypeError that aborts init and leaves the page stuck on the 'Assembling jeweled movement...' loading overlay forever. Cause: the model passed the tube segment count as a continuous value without rounding. Changing it to Math.round(turns * 38) makes the count an integer, so the scene loads and the overlay hides on schedule. Bug 2: an abnormally blinding glow fills the center of the scene, washing the whole dial to pure white so nothing is legible. Cause: UnrealBloomPass (line 340) used too low a threshold (0.18) and a high strength (0.72), which together with ACES tone mapping and several high-intensity lights pulled nearly every metallic highlight and lume surface into the bloom and spread it into one big over-exposed white blob. To protect testers' eyes, the bloom was tamed to strength 0.28 / radius 0.45 / threshold 0.9: the glow now only haloes the lume markers softly, and the movement, indices and hands are legible again."
},
"contributor": "Nagi-ovo"
}
}
}