INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.13

2:0.14

3:0.18

4:0.12

5:0.09

6:0.10

7:0.13

Negative Logits

-3.70

-2.84

in

-2.55

↵↵

-2.53

-2.48

-2.42

-2.39

<eos>

-2.31

and

-2.28

-2.27

POSITIVE LOGITS

 myſelf

5.69

 itſelf

5.50

Efq

5.41

 Monfieur

5.28

 Jefus

5.03

 Theſe

5.03

 ―――――

5.00

 pleaſure

4.94

 Anſ

4.88

 Majefty

4.84

Activations Density 0.000%

No Known Activations

This feature has no known activations.