Correct Key Results table + add FPR explanation

README.md

@@ -43,11 +43,17 @@ A 184M-parameter DeBERTa-v3-base guardrail classifier trained on 57,000+ real-wo
   - B-10 Regulatory Enquiry Mishandling — EU AI Act Art. 52
   - B-11 AML/Sanctions Evasion — FATF Recommendation 10
 
-**Results vs LlamaGuard-3-8B across 22 benchmarks:**
-- Wins all 7 prompt-injection benchmarks
-- 0% false positive rate on 208 agentic tasks (vs 6.3% for LlamaGuard-3-8B)
-- 11.6ms inference latency — 44× fewer parameters
-- Deployable as always-on inline guardrail without GPU infrastructure
+---
+
+## Key Results
+
+| Model | Size | HackaPrompt R | AgentHarm FPR | WildGuardMix F1 | Latency |
+|---|---|---|---|---|---|
+| **Semalith v1.5** | **184M** | **0.994** | **0.5%** | **0.303** | **11.6ms** |
+| PromptGuard-86M | 86M | 1.000 | 96.9% | 0.095 | 8ms |
+| LlamaGuard-3-8B | 8B | — | — | — | ~180ms |
+
+**FPR (False Positive Rate)** measures how often a model incorrectly flags a legitimate, benign request as an attack. A model with high FPR cannot be deployed in production — it blocks real customers. PromptGuard's 96.9% FPR means it flags nearly every legitimate agentic task as an injection attack, making it a syntax detector rather than a safety classifier. Semalith's 0.5% FPR (1 false alarm in 208 agentic tasks) reflects domain-specific training that distinguishes attack intent from legitimate BFSI queries.
 
 ---
 
@@ -75,18 +81,6 @@ Generated using **Quantum Circuit Born Machine (QCBM)** sampling on PennyLane
 
 1,273 adversarial prompts generated via QCBM + simulated annealing targeting Semalith's decision boundary. Covers professional and retail registers. Overall Semalith miss rate: 14.3%.
 
-Techniques: SA Annealing (344), QCBM 8q boundary (173), QCBM 8q gradient (125), 10q Paraphrase Fix1 (123), QCBM 8q joint (100), retail customer mobile (157), RM internal (105), PG-miss professional (84), PG adversarial B-03 (3).
-
----
-
-## Key Results
-
-| Model | Size | HackaPrompt R | AgentHarm FPR | WildGuardMix F1 | Latency |
-|---|---|---|---|---|---|
-| **Semalith v1.5** | **184M** | **0.994** | **0.000** | **0.62** | **11.6ms** |
-| LlamaGuard-3-8B | 8B | 0.941 | 0.063 | 0.58 | ~180ms |
-| PromptGuard-86M | 86M | 0.981 | 0.126 | 0.41 | 8ms |
-
 ---
 
 ## Research