feat: update .gitignore, add experiment logs and results for naive backend

Author: Neal006

.gitignore

@@ -1,4 +1,5 @@
 .env
+.claude/
 __pycache__/
 *.pyc
 *.pyo

README.md

@@ -13,7 +13,7 @@
 [![Stars](https://img.shields.io/github/stars/Neal006/memorylens?style=social)](https://github.com/Neal006/memorylens/stargazers)
 [![Forks](https://img.shields.io/github/forks/Neal006/memorylens?style=social)](https://github.com/Neal006/memorylens/network/members)
 
-[**Quick Start**](#quick-start) · [**Results**](#benchmark-results) · [**How It Works**](#how-it-works) · [**vs Other Tools**](#how-memorylens-compares) · [**Contributing**](#contributing) · [**Paper**](paper/memorylens_paper.md)
+[**Quick Start**](#quick-start) · [**Results**](#benchmark-results) · [**How It Works**](#how-it-works) · [**vs Other Tools**](#how-memorylens-compares)
 
 </div>
 
@@ -45,7 +45,7 @@ Run `python main.py` and get statistically valid results like these — **no API
 | **Cascading Temporal** (Ebbinghaus decay) | **87.5 ± 0.0%** | **218** | **5.67×** |
 | SummaryMemory (rolling compression) | 100.0 ± 0.0% | 318 | — |
 
-> **Chunked RAG vs Ideal RAG** shows the gap between a theoretical upper bound and a production-realistic retrieval system. The 15pp difference is what chunking + index eviction costs you. The **Cascading Temporal** backend delivers **5.67× more recall per token** than naive truncation using an Ebbinghaus-grounded forgetting curve — now cited and ablated in the [research paper](paper/memorylens_paper.md).
+> **Chunked RAG vs Ideal RAG** shows the gap between a theoretical upper bound and a production-realistic retrieval system. The 15pp difference is what chunking + index eviction costs you. The **Cascading Temporal** backend delivers **5.67× more recall per token** than naive truncation using an Ebbinghaus-grounded forgetting curve.
 
 ---
 

experiment_logs/runs_summary.csv

@@ -0,0 +1,3 @@
+run_id,backend,turn,recall,precision,drift,noise,tokens,total_turns
+test_run,naive,10,1.0,0.9,0.0,0.5,100,25
+test_run,naive,25,0.8,0.7,0.1,0.8,500,25

experiment_logs/test_run.json

@@ -0,0 +1,37 @@
+{
+  "run_id": "test_run",
+  "config": {
+    "total_turns": 25,
+    "backends": [
+      "naive"
+    ]
+  },
+  "results": {
+    "checkpoints": [
+      10,
+      25
+    ],
+    "naive": {
+      "recall": [
+        1.0,
+        0.8
+      ],
+      "precision": [
+        0.9,
+        0.7
+      ],
+      "drift": [
+        0.0,
+        0.1
+      ],
+      "noise": [
+        0.5,
+        0.8
+      ],
+      "tokens": [
+        100,
+        500
+      ]
+    }
+  }
+}