INDEX

Explanations

attends to the token "which" from the token "to."

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

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.89

1:0.01

2:0.01

3:0.01

4:0.02

5:0.01

6:0.01

7:0.02

Negative Logits

he

-0.53

we

-0.52

-0.46

be

-0.44

is

-0.42

 will

-0.40

 they

-0.39

 must

-0.38

are

-0.38

POSITIVE LOGITS

 Signalez

0.58

 disambiguazione

0.55

出版年

0.53

UserScript

0.52

 ModelExpression

0.52

TagMode

0.50

Datuak

0.50

########.

0.49

 виправивши

0.49

 propOrder

0.48

Activations Density 0.240%

attends to the token "which" from the token "to."

No Comments

No Known Activations