Skip to content

Commit e207dc8

Browse files
committed
readme: move how-it-works above the problem
1 parent 4de79e7 commit e207dc8

1 file changed

Lines changed: 6 additions & 6 deletions

File tree

README.md

Lines changed: 6 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -19,6 +19,12 @@
1919

2020
---
2121

22+
## How it works
23+
24+
<p align="center">
25+
<img src="docs/how-it-works.png" width="700" alt="How IntentProbe works: tool description → frozen model → slice open layers 13-15 → 22KB probe → allow/warn/block" />
26+
</p>
27+
2228
## The problem
2329

2430
```
@@ -102,12 +108,6 @@ Head-to-head, same test sets, same split, same seed. Every number is reproducibl
102108
>
103109
> Five rounds of experiments, each removing a text-level shortcut. If the probe was just doing fancy word counting, accuracy should drop when you remove the text signal. It never did. TF-IDF went from 93% to 30%. The activation probe stayed above 93% throughout. The activations contain a signal that text surfaces don't.
104110
105-
## How it works
106-
107-
<p align="center">
108-
<img src="docs/how-it-works.png" width="700" alt="How IntentProbe works: tool description → frozen model → slice open layers 13-15 → 22KB probe → allow/warn/block" />
109-
</p>
110-
111111
## Install
112112

113113
```bash

0 commit comments

Comments
 (0)