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
57 lines
9.4 KiB
XML
57 lines
9.4 KiB
XML
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 40 880 500" width="880" height="500" 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="120" y="55" width="640" height="35" rx="4" fill="#f5f5f5" stroke="#999999" stroke-width="2"/>
|
|
<text x="135" y="73" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">Task: Xfinity plan change | Issue: Missing billing address + Incorrect plan name</text>
|
|
<rect x="37.5" y="100" width="190" height="36" rx="6" fill="#999999" stroke="#333333" stroke-width="2"/>
|
|
<text x="132.5" y="118" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#ffffff" text-anchor="middle" dominant-baseline="central" font-weight="bold">Scalar</text>
|
|
<rect x="37.5" y="146" width="190" height="210" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
|
<text x="47.5" y="165" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central"></text>
|
|
<text x="47.5" y="191" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central"> 7.2 points</text>
|
|
<text x="47.5" y="217" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central"></text>
|
|
<text x="47.5" y="243" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">─────────</text>
|
|
<text x="47.5" y="269" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">No diagnostic capability</text>
|
|
<text x="47.5" y="295" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">Don't know where</text>
|
|
<text x="47.5" y="321" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">Good/bad</text>
|
|
<line x1="229.5" y1="250" x2="240.5" y2="250" stroke="#333333" stroke-width="2" marker-end="url(#ah)"/>
|
|
<rect x="242.5" y="100" width="190" height="36" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
|
<text x="337.5" y="118" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#ffffff" text-anchor="middle" dominant-baseline="central" font-weight="bold">Semi-scalar</text>
|
|
<rect x="242.5" y="146" width="190" height="210" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
|
<text x="252.5" y="165" font-family="'Courier New', Courier, monospace" font-size="9" fill="#333333" text-anchor="start" dominant-baseline="central">Good: Found correct phone number</text>
|
|
<text x="252.5" y="191" font-family="'Courier New', Courier, monospace" font-size="7.5" fill="#333333" text-anchor="start" dominant-baseline="central">Good: Successfully completed the change</text>
|
|
<text x="252.5" y="217" font-family="'Courier New', Courier, monospace" font-size="10" fill="#333333" text-anchor="start" dominant-baseline="central">Bad: Missing billing address</text>
|
|
<text x="252.5" y="243" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">Bad: Incorrect plan name</text>
|
|
<text x="252.5" y="269" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">─────────</text>
|
|
<text x="252.5" y="295" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">Overall: 6.5 points</text>
|
|
<text x="252.5" y="321" font-family="'Courier New', Courier, monospace" font-size="8" fill="#333333" text-anchor="start" dominant-baseline="central">(Evidence-based but coarse-grained)</text>
|
|
<line x1="434.5" y1="250" x2="445.5" y2="250" stroke="#333333" stroke-width="2" marker-end="url(#ah)"/>
|
|
<rect x="447.5" y="100" width="190" height="36" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
|
<text x="542.5" y="118" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#ffffff" text-anchor="middle" dominant-baseline="central" font-weight="bold">Vector</text>
|
|
<rect x="447.5" y="146" width="190" height="230" rx="6" fill="#f0f0f0" stroke="#333333" stroke-width="2"/>
|
|
<text x="457.5" y="165" font-family="'Courier New', Courier, monospace" font-size="9" fill="#333333" text-anchor="start" dominant-baseline="central">Information query accuracy 9/10</text>
|
|
<text x="457.5" y="191" font-family="'Courier New', Courier, monospace" font-size="7" fill="#333333" text-anchor="start" dominant-baseline="central">Information collection completeness 6/10</text>
|
|
<text x="457.5" y="217" font-family="'Courier New', Courier, monospace" font-size="11" fill="#333333" text-anchor="start" dominant-baseline="central">Communication fluency 8/10</text>
|
|
<text x="457.5" y="243" font-family="'Courier New', Courier, monospace" font-size="10.5" fill="#333333" text-anchor="start" dominant-baseline="central">Communication accuracy 7/10</text>
|
|
<text x="457.5" y="269" font-family="'Courier New', Courier, monospace" font-size="8.5" fill="#333333" text-anchor="start" dominant-baseline="central">User communication accuracy 10/10</text>
|
|
<text x="457.5" y="295" font-family="'Courier New', Courier, monospace" font-size="10" fill="#333333" text-anchor="start" dominant-baseline="central">Overall task completion 8/10</text>
|
|
<text x="457.5" y="321" font-family="'Courier New', Courier, monospace" font-size="12" fill="#333333" text-anchor="start" dominant-baseline="central">─────────</text>
|
|
<text x="457.5" y="343" font-family="'Courier New', Courier, monospace" font-size="9" fill="#333333" text-anchor="start" dominant-baseline="central">Precisely locate shortcomings ↑</text>
|
|
<line x1="639.5" y1="250" x2="650.5" y2="250" stroke="#333333" stroke-width="2" marker-end="url(#ah)"/>
|
|
<rect x="652.5" y="100" width="190" height="36" rx="6" fill="#999999" stroke="#333333" stroke-width="2"/>
|
|
<text x="747.5" y="118" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#ffffff" text-anchor="middle" dominant-baseline="central" font-weight="bold">Generative</text>
|
|
<rect x="652.5" y="146" width="190" height="210" rx="6" fill="#d0d0d0" stroke="#333333" stroke-width="2"/>
|
|
<text x="662.5" y="165" font-family="'Courier New', Courier, monospace" font-size="9" fill="#333333" text-anchor="start" dominant-baseline="central">Perspective 1: Accurate tool use</text>
|
|
<text x="662.5" y="191" font-family="'Courier New', Courier, monospace" font-size="6.5" fill="#333333" text-anchor="start" dominant-baseline="central">Perspective 2: Missing information collection</text>
|
|
<text x="662.5" y="217" font-family="'Courier New', Courier, monospace" font-size="7.5" fill="#333333" text-anchor="start" dominant-baseline="central">Perspective 3: Plan name paraphrasing</text>
|
|
<text x="662.5" y="243" font-family="'Courier New', Courier, monospace" font-size="8" fill="#333333" text-anchor="start" dominant-baseline="central"> Error → Suggest direct repetition</text>
|
|
<text x="662.5" y="269" font-family="'Courier New', Courier, monospace" font-size="6" fill="#333333" text-anchor="start" dominant-baseline="central">Perspective 4: Not explaining information usage</text>
|
|
<text x="662.5" y="295" font-family="'Courier New', Courier, monospace" font-size="6.5" fill="#333333" text-anchor="start" dominant-baseline="central">Perspective 5: Not verifying agent identity</text>
|
|
<text x="662.5" y="321" font-family="'Courier New', Courier, monospace" font-size="10.5" fill="#333333" text-anchor="start" dominant-baseline="central">→ 6-7 analysis perspectives</text>
|
|
<line x1="132.5" y1="368" x2="747.5" y2="368" stroke="#333333" stroke-width="2" marker-end="url(#ah)"/>
|
|
<text x="440.0" y="388" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="16" fill="#666666" text-anchor="middle" dominant-baseline="central" font-weight="bold">Increasing information →</text>
|
|
<rect x="60" y="405" width="760" height="55" rx="4" fill="#f5f5f5" stroke="#999999" stroke-width="2"/>
|
|
<text x="80" y="420" font-family="'Courier New', Courier, monospace" font-size="15" fill="#333333" text-anchor="start" dominant-baseline="central">DeepSeek experiment: Generative 8 evaluations produce 6-7 analytical perspectives</text>
|
|
<text x="80" y="442" font-family="'Courier New', Courier, monospace" font-size="9" fill="#333333" text-anchor="start" dominant-baseline="central">Significant improvement after comprehensive multi-angle diagnosis (illustrative) — learning improvement direction from a single failure</text>
|
|
<rect x="60" y="475" width="760" height="45" rx="6" fill="#999999" stroke="#333333" stroke-width="2"/>
|
|
<text x="440.0" y="498" font-family="Arial, 'Helvetica Neue', Helvetica, 'PingFang SC', 'Microsoft YaHei', sans-serif" font-size="13" fill="#ffffff" text-anchor="middle" dominant-baseline="central" font-weight="bold">Generative reward: transforming rich environmental feedback into learnable knowledge without hundreds of trials and errors</text>
|
|
</svg> |