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

0:0.23

1:0.06

2:0.07

3:0.14

4:0.12

5:0.12

6:0.15

7:0.07

Negative Logits

 itſelf

-1.55

 ſche

-1.49

 myſelf

-1.48

 Monfieur

-1.48

 ſtate

-1.47

 Reſ

-1.47

 ſever

-1.44

 houſe

-1.42

 pleaſure

-1.42

 Diſ

-1.41

POSITIVE LOGITS

0.97

so

0.89

in

0.88

the

0.87

an

0.85

0.83

for

0.82

“

0.78

and

0.77

Activations Density 0.000%

No Known Activations

This feature has no known activations.