RohitChandramouli6618 commited on
Commit
a251104
·
1 Parent(s): 7c731ea

Update greedy baseline numbers from latest test_local run; recalculate lifts

Browse files
Files changed (3) hide show
  1. README.md +3 -3
  2. config/tasks.yaml +2 -2
  3. server/app.py +21 -21
README.md CHANGED
@@ -214,9 +214,9 @@ Results from `baseline/run.py` (Llama 3.3 70B via Groq, runtime ~19.8 minutes):
214
  | Task | Greedy (D0) | LLM+GRPO | Lift |
215
  | --- | --- | --- | --- |
216
  | Easy | 42.8% (breach 60%) | **90.8%** | +48pp |
217
- | Medium | 42.7% (breach 80%) | **78.0%** | +35pp |
218
- | Hard | 33.0% (breach 100%) | **61.1%** | +28pp |
219
- | **Average** | **39.5%** | **75.9%** | **+36pp** |
220
 
221
  The greedy baseline (always allocates to D0) scores 33–43% with 60–100% hospital breach rates — no trivial exploit path. LLM+GRPO reaches 66–91% with zero breaches, demonstrating that genuine triage reasoning is required and rewarded.
222
 
 
214
  | Task | Greedy (D0) | LLM+GRPO | Lift |
215
  | --- | --- | --- | --- |
216
  | Easy | 42.8% (breach 60%) | **90.8%** | +48pp |
217
+ | Medium | 39.6% (breach 100%) | **78.0%** | +38pp |
218
+ | Hard | 35.3% (breach 100%) | **61.1%** | +26pp |
219
+ | **Average** | **39.2%** | **75.9%** | **+37pp** |
220
 
221
  The greedy baseline (always allocates to D0) scores 33–43% with 60–100% hospital breach rates — no trivial exploit path. LLM+GRPO reaches 66–91% with zero breaches, demonstrating that genuine triage reasoning is required and rewarded.
222
 
config/tasks.yaml CHANGED
@@ -81,8 +81,8 @@ benchmark:
81
  hard: 3
82
  greedy_d0:
83
  easy: {score: 0.428, breach_rate: 0.60}
84
- medium: {score: 0.427, breach_rate: 0.80}
85
- hard: {score: 0.330, breach_rate: 1.00}
86
  llm_grpo:
87
  easy: {score: 0.885, model: llama-3.1-8b-instant}
88
  medium: {score: 0.766, model: llama-3.1-8b-instant}
 
81
  hard: 3
82
  greedy_d0:
83
  easy: {score: 0.428, breach_rate: 0.60}
84
+ medium: {score: 0.396, breach_rate: 1.00}
85
+ hard: {score: 0.353, breach_rate: 1.00}
86
  llm_grpo:
87
  easy: {score: 0.885, model: llama-3.1-8b-instant}
88
  medium: {score: 0.766, model: llama-3.1-8b-instant}
server/app.py CHANGED
@@ -569,7 +569,7 @@ body{font-family:var(--mono);background:var(--bg);color:var(--text);min-height:1
569
  <div class="hero-stats">
570
  <div class="hs-item"><div class="hs-val">3</div><div class="hs-label">Task levels</div></div>
571
  <div class="hs-item"><div class="hs-val">75.9%</div><div class="hs-label">LLM+GRPO avg</div></div>
572
- <div class="hs-item"><div class="hs-val">+36pp</div><div class="hs-label">vs greedy</div></div>
573
  </div>
574
  </div>
575
 
@@ -729,21 +729,21 @@ body{font-family:var(--mono);background:var(--bg);color:var(--text);min-height:1
729
  </div>
730
  <div class="card-sm" style="border-top:2px solid var(--amber);">
731
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--amber);margin-bottom:0.5rem;">Medium</div>
732
- <div style="font-family:var(--serif);font-size:2rem;font-weight:700;color:#fff;line-height:1;">42.7%</div>
733
- <div style="font-size:0.68rem;color:var(--muted);margin-top:0.25rem;">&sigma;=0.059 &middot; breach 80%</div>
734
  <div style="margin-top:0.75rem;display:flex;flex-direction:column;gap:0.3rem;">
735
- <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Containment</span><span>45%</span></div>
736
- <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Hospital</span><span>59%</span></div>
737
- <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Efficiency</span><span>8%</span></div>
738
  </div>
739
  </div>
740
  <div class="card-sm" style="border-top:2px solid var(--red);">
741
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--red);margin-bottom:0.5rem;">Hard</div>
742
- <div style="font-family:var(--serif);font-size:2rem;font-weight:700;color:#fff;line-height:1;">33.0%</div>
743
  <div style="font-size:0.68rem;color:var(--muted);margin-top:0.25rem;">&sigma;=0.020 &middot; breach 100%</div>
744
  <div style="margin-top:0.75rem;display:flex;flex-direction:column;gap:0.3rem;">
745
- <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Containment</span><span>30%</span></div>
746
- <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Hospital</span><span>50%</span></div>
747
  <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Efficiency</span><span style="color:var(--red);">0%</span></div>
748
  </div>
749
  </div>
@@ -849,22 +849,22 @@ body{font-family:var(--mono);background:var(--bg);color:var(--text);min-height:1
849
  </tr>
850
  <tr>
851
  <td><span class="badge badge-amber">Medium</span></td>
852
- <td style="color:var(--muted);">42.7%</td><td><strong>78.0%</strong></td>
853
- <td class="pos">+35pp</td>
854
  <td><span class="badge badge-green">Strong</span></td>
855
  <td><span class="badge badge-green">None &mdash; 8% eff, 80% breach</span></td>
856
  </tr>
857
  <tr>
858
  <td><span class="badge badge-red">Hard</span></td>
859
- <td style="color:var(--muted);">33.0%</td><td><strong>61.1%</strong></td>
860
- <td class="pos">+28pp</td>
861
  <td><span class="badge badge-green">Strong</span></td>
862
  <td><span class="badge badge-green">None &mdash; 0% eff, 100% breach</span></td>
863
  </tr>
864
  <tr style="border-top:1px solid var(--border2);">
865
  <td><strong>Average</strong></td>
866
- <td style="color:var(--muted);">39.5%</td><td><strong>75.9%</strong></td>
867
- <td class="pos"><strong>+36pp</strong></td>
868
  <td><span class="badge badge-green">Strong</span></td>
869
  <td><span class="badge badge-green">No exploits found</span></td>
870
  </tr>
@@ -874,7 +874,7 @@ body{font-family:var(--mono);background:var(--bg);color:var(--text);min-height:1
874
  <div class="card-sm">
875
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Easy</div>
876
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
877
- <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>43%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:43%;background:var(--dim);"></div></div></div>
878
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--green);">LLM+GRPO</span><span>89%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:89%;background:var(--green);"></div></div></div>
879
  <div style="font-size:0.68rem;color:var(--muted);">Lift: <span style="color:var(--green);font-weight:700;">+46pp</span></div>
880
  </div>
@@ -882,22 +882,22 @@ body{font-family:var(--mono);background:var(--bg);color:var(--text);min-height:1
882
  <div class="card-sm">
883
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Medium</div>
884
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
885
- <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>43%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:43%;background:var(--dim);"></div></div></div>
886
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--amber);">LLM+GRPO</span><span>75%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:75%;background:var(--amber);"></div></div></div>
887
- <div style="font-size:0.68rem;color:var(--muted);">Lift: <span style="color:var(--amber);font-weight:700;">+35pp</span></div>
888
  </div>
889
  </div>
890
  <div class="card-sm">
891
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Hard</div>
892
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
893
- <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>33%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:33%;background:var(--dim);"></div></div></div>
894
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--red);">LLM+GRPO</span><span>63%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:63%;background:var(--red);"></div></div></div>
895
- <div style="font-size:0.68rem;color:var(--muted);">Lift: <span style="color:var(--red);font-weight:700;">+28pp</span></div>
896
  </div>
897
  </div>
898
  </div>
899
  <div style="padding:0.85rem 1.1rem;background:var(--green-dim);border:1px solid rgba(61,214,140,0.2);border-radius:8px;font-size:0.72rem;color:var(--text);line-height:1.8;">
900
- &#10003; <strong>Variance check passed.</strong> Mean lift of <strong>+36pp</strong> across all tasks.
901
  Greedy (D0) scores 33&ndash;43% with 60&ndash;100% hospital breach rates &mdash; confirming no trivial exploit path.
902
  LLM+GRPO reaches 66&ndash;91% with zero breaches &mdash; genuine triage reasoning is required and rewarded.
903
  </div>
 
569
  <div class="hero-stats">
570
  <div class="hs-item"><div class="hs-val">3</div><div class="hs-label">Task levels</div></div>
571
  <div class="hs-item"><div class="hs-val">75.9%</div><div class="hs-label">LLM+GRPO avg</div></div>
572
+ <div class="hs-item"><div class="hs-val">+37pp</div><div class="hs-label">vs greedy</div></div>
573
  </div>
574
  </div>
575
 
 
729
  </div>
730
  <div class="card-sm" style="border-top:2px solid var(--amber);">
731
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--amber);margin-bottom:0.5rem;">Medium</div>
732
+ <div style="font-family:var(--serif);font-size:2rem;font-weight:700;color:#fff;line-height:1;">39.6%</div>
733
+ <div style="font-size:0.68rem;color:var(--muted);margin-top:0.25rem;">&sigma;=0.026 &middot; breach 100%</div>
734
  <div style="margin-top:0.75rem;display:flex;flex-direction:column;gap:0.3rem;">
735
+ <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Containment</span><span>42%</span></div>
736
+ <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Hospital</span><span>52%</span></div>
737
+ <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Efficiency</span><span>9%</span></div>
738
  </div>
739
  </div>
740
  <div class="card-sm" style="border-top:2px solid var(--red);">
741
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--red);margin-bottom:0.5rem;">Hard</div>
742
+ <div style="font-family:var(--serif);font-size:2rem;font-weight:700;color:#fff;line-height:1;">35.3%</div>
743
  <div style="font-size:0.68rem;color:var(--muted);margin-top:0.25rem;">&sigma;=0.020 &middot; breach 100%</div>
744
  <div style="margin-top:0.75rem;display:flex;flex-direction:column;gap:0.3rem;">
745
+ <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Containment</span><span>35%</span></div>
746
+ <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Hospital</span><span>52%</span></div>
747
  <div style="display:flex;justify-content:space-between;font-size:0.68rem;"><span style="color:var(--muted);">Efficiency</span><span style="color:var(--red);">0%</span></div>
748
  </div>
749
  </div>
 
849
  </tr>
850
  <tr>
851
  <td><span class="badge badge-amber">Medium</span></td>
852
+ <td style="color:var(--muted);">39.6%</td><td><strong>78.0%</strong></td>
853
+ <td class="pos">+38pp</td>
854
  <td><span class="badge badge-green">Strong</span></td>
855
  <td><span class="badge badge-green">None &mdash; 8% eff, 80% breach</span></td>
856
  </tr>
857
  <tr>
858
  <td><span class="badge badge-red">Hard</span></td>
859
+ <td style="color:var(--muted);">35.3%</td><td><strong>61.1%</strong></td>
860
+ <td class="pos">+26pp</td>
861
  <td><span class="badge badge-green">Strong</span></td>
862
  <td><span class="badge badge-green">None &mdash; 0% eff, 100% breach</span></td>
863
  </tr>
864
  <tr style="border-top:1px solid var(--border2);">
865
  <td><strong>Average</strong></td>
866
+ <td style="color:var(--muted);">39.2%</td><td><strong>75.9%</strong></td>
867
+ <td class="pos"><strong>+37pp</strong></td>
868
  <td><span class="badge badge-green">Strong</span></td>
869
  <td><span class="badge badge-green">No exploits found</span></td>
870
  </tr>
 
874
  <div class="card-sm">
875
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Easy</div>
876
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
877
+ <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>40%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:40%;background:var(--dim);"></div></div></div>
878
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--green);">LLM+GRPO</span><span>89%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:89%;background:var(--green);"></div></div></div>
879
  <div style="font-size:0.68rem;color:var(--muted);">Lift: <span style="color:var(--green);font-weight:700;">+46pp</span></div>
880
  </div>
 
882
  <div class="card-sm">
883
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Medium</div>
884
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
885
+ <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>40%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:40%;background:var(--dim);"></div></div></div>
886
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--amber);">LLM+GRPO</span><span>75%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:75%;background:var(--amber);"></div></div></div>
887
+ <div style="font-size:0.68rem;color:var(--muted);">Lift: <span style="color:var(--amber);font-weight:700;">+38pp</span></div>
888
  </div>
889
  </div>
890
  <div class="card-sm">
891
  <div style="font-size:0.62rem;letter-spacing:0.08em;text-transform:uppercase;color:var(--muted);margin-bottom:0.75rem;">Hard</div>
892
  <div style="display:flex;flex-direction:column;gap:0.45rem;">
893
+ <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--muted);">Greedy (D0)</span><span>35%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:35%;background:var(--dim);"></div></div></div>
894
  <div><div style="display:flex;justify-content:space-between;font-size:0.68rem;margin-bottom:0.2rem;"><span style="color:var(--red);">LLM+GRPO</span><span>63%</span></div><div class="bar-track" style="height:8px;"><div class="bar-fill" style="width:63%;background:var(--red);"></div></div></div>
895
+ <div style="font-size:0.68rem;color:var(--red);font-weight:700;">+26pp</span></div>
896
  </div>
897
  </div>
898
  </div>
899
  <div style="padding:0.85rem 1.1rem;background:var(--green-dim);border:1px solid rgba(61,214,140,0.2);border-radius:8px;font-size:0.72rem;color:var(--text);line-height:1.8;">
900
+ &#10003; <strong>Variance check passed.</strong> Mean lift of <strong>+37pp</strong> across all tasks.
901
  Greedy (D0) scores 33&ndash;43% with 60&ndash;100% hospital breach rates &mdash; confirming no trivial exploit path.
902
  LLM+GRPO reaches 66&ndash;91% with zero breaches &mdash; genuine triage reasoning is required and rewarded.
903
  </div>