liqiang
b119135836
Build latest book artifacts / build (push) Canceled after 0s
dependency resolution / resolve (3.11) (push) Canceled after 0s
dependency resolution / resolve (3.13) (push) Canceled after 0s
deploy-pages / build (push) Canceled after 0s
deploy-pages / deploy (push) Canceled after 0s
i18n consistency check / check (push) Canceled after 0s
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
web-search-agent tests / test (push) Canceled after 0s
web-search-agent tests / agentbook (push) Canceled after 0s
51 lines
11 KiB
XML
51 lines
11 KiB
XML
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 40 780 490" width="780" height="490" style="background:#ffffff">
|
||
<defs><marker id="ah" markerWidth="12" markerHeight="8" refX="12" refY="4" orient="auto"><polygon points="0 0, 12 4, 0 8" fill="#333333" /></marker><marker id="ah-light" markerWidth="12" markerHeight="8" refX="12" refY="4" orient="auto"><polygon points="0 0, 12 4, 0 8" fill="#999999" /></marker></defs>
|
||
|
||
<rect x="30" y="55" width="340" height="165" rx="8" fill="#ffffff" stroke="#333333" stroke-width="2" stroke-dasharray="8,4" />
|
||
<text x="42" y="137.5" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="bold">GP-L (văn bản thuần túy)</text>
|
||
<rect x="45" y="82" width="150" height="50" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2" />
|
||
<text x="120" y="100" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">tập huấn luyện</text>
|
||
<text x="120" y="116" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">J/Q/K = 10</text>
|
||
<line x1="197" y1="107" x2="218" y2="107" stroke="#333333" stroke-width="2" marker-end="url(#ah)" />
|
||
<rect x="220" y="82" width="135" height="50" rx="4" fill="#f5f5f5" stroke="#333333" stroke-width="2" />
|
||
<text x="287" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="12.88" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold"><tspan x="287" dy="93.96">Bộ kiểm tra</tspan><tspan x="287" dy="13.91">(OOD)</tspan></text>
|
||
<text x="287" y="120.91" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.27" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">J=11, Q=12, K=13</text>
|
||
<text x="195" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="195" dy="114.82">Đầu vào: "Tính 24 bằng cách sử dụng [7, J, 3,</tspan><tspan x="195" dy="15.12">5]"</tspan></text>
|
||
<text x="195" y="145.06" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">Đầu ra: "(J-7)×(5+3) = 3×8 = 24"</text>
|
||
<text x="195" y="160.18" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">* Khi OOD J=11: "(11-7)×(5+3)=32≠24"</text>
|
||
<rect x="410" y="55" width="340" height="165" rx="8" fill="#ffffff" stroke="#333333" stroke-width="2" stroke-dasharray="8,4" />
|
||
<text x="422" y="137.5" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#666666" text-anchor="start" dominant-baseline="central" font-weight="bold">GP-VL (Ngôn ngữ hình ảnh)</text>
|
||
<rect x="425" y="82" width="150" height="50" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2" />
|
||
<text x="500" y="100" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">tập huấn luyện</text>
|
||
<text x="500" y="116" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">♠♣ màu đen</text>
|
||
<line x1="577" y1="107" x2="598" y2="107" stroke="#333333" stroke-width="2" marker-end="url(#ah)" />
|
||
<rect x="600" y="82" width="135" height="50" rx="4" fill="#f5f5f5" stroke="#333333" stroke-width="2" />
|
||
<text x="667" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="12.88" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold"><tspan x="667" dy="93.96">Bộ kiểm tra</tspan><tspan x="667" dy="13.91">(OOD)</tspan></text>
|
||
<text x="667" y="120.91" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.27" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">♥♦ màu đỏ</text>
|
||
<text x="575" y="114.82" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">Đầu vào: hình ảnh thẻ chơi</text>
|
||
<text x="575" y="129.94" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">Bộ mã hóa hình ảnh → Nhận dạng số → Tư duy</text>
|
||
<text x="575" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="575" dy="145.06">* OOD: Sự thay đổi màu sắc bề ngoài kiểm tra</tspan><tspan x="575" dy="15.12">độ chắc chắn của thị giác</tspan></text>
|
||
<text x="390" y="245" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="20" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Llama-3.2-Vision-11B</text>
|
||
<rect x="40" y="265" width="320" height="50" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2" />
|
||
<text x="200" y="282" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Đường dẫn SFT</text>
|
||
<text x="200" y="302" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">max Σ log P(y|x) | Phân bổ đào tạo trí nhớ</text>
|
||
<rect x="420" y="265" width="320" height="50" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2" />
|
||
<text x="580" y="277.28" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13.47" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Đường dẫn RL (PPO)</text>
|
||
<text x="580" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.78" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="580" dy="290.91">max E[R(τ)] | Khám phá các chiến lược tổng</tspan><tspan x="580" dy="12.72">quát hóa</tspan></text>
|
||
<rect x="30" y="340" width="720" height="160" rx="4" fill="#f5f5f5" stroke="#333333" stroke-width="2" />
|
||
<text x="390" y="339" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="11.87" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Kết quả kiểm tra OOD (thay đổi từ đường cơ sở SFT Init)</text>
|
||
<text x="100" y="350.81" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold" />
|
||
<text x="245" y="361.61" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Quy tắc OOD (GP-L)</text>
|
||
<text x="415" y="372.41" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Quy tắc OOD (GP-VL)</text>
|
||
<text x="590" y="404.88" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">Tầm nhìn OOD (GP-VL)</text>
|
||
<text x="100" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="100" dy="383.21">Tiện ích mở rộng</tspan><tspan x="100" dy="10.8">SFT</tspan></text>
|
||
<text x="245" y="404.81" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">−8.1% (11.5→3.4)</text>
|
||
<text x="415" y="415.61" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">−5.6%</text>
|
||
<text x="590" y="420" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal">−9.9% (23.6→13.7)</text>
|
||
<text x="100" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="100" dy="426.41">Tiện ích mở rộng</tspan><tspan x="100" dy="10.8">RL</tspan></text>
|
||
<text x="245" y="448.01" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">+3.5% (11.5→15.0)</text>
|
||
<text x="415" y="458.81" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">+3.0%</text>
|
||
<text x="590" y="435.12" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="14" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold">+17.6% (23.6→41.2)</text>
|
||
<text x="390" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#333333" text-anchor="middle" dominant-baseline="central" font-weight="bold"><tspan x="390" dy="469.61">Kết luận: SFT giảm đồng đều bên ngoài phân phối, RL cải thiện đồng đều - "Bộ nhớ</tspan><tspan x="390" dy="10.8">SFT, tổng quát hóa RL"</tspan></text>
|
||
<text x="390" y="0" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="10" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="normal"><tspan x="390" dy="491.21">Tiền đề chính: RL end-to-end không có SFT thất bại hoàn toàn (đầu ra không có cấu trúc, không</tspan><tspan x="390" dy="10.8">thể tính phần thưởng)</tspan></text>
|
||
</svg> |
