INDEX

Explanations

attends to consecutive code or structural modifications marked by specific changes from one component to another

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.09

1:0.09

2:0.15

3:0.11

4:0.07

5:0.04

6:0.18

7:0.23

Negative Logits

expandindo

-0.37

MigrationBuilder

-0.36

stdarg

-0.34

?',

-0.33

Története

-0.33

Хьажоргаш

-0.32

{}'.

-0.31

лтамалар

-0.31

 '\\;'

-0.30

خواندن

-0.30

POSITIVE LOGITS

Hentet

0.28

<caption>

0.27

 Económica

0.26

Mee

0.26

umente

0.26

dib

0.25

greg

0.25

#+#

0.25

oplast

0.25

gliere

0.25

Activations Density 0.030%

attends to consecutive code or structural modifications marked by specific changes from one component to another

No Comments

No Known Activations