INDEX

Explanations

attends to the token "and" from any preceding token pattern

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

0:0.56

1:0.06

2:0.05

3:0.03

4:0.08

5:0.06

6:0.05

7:0.07

Negative Logits

-0.44

{}/

-0.44

</tfoot>

-0.43

rouvez

-0.38

ratulations

-0.37

bibfield

-0.37

:]:

-0.36

Pozri

-0.35

BibitemShut

-0.35

ustimmung

-0.35

POSITIVE LOGITS

 universitario

0.49

 esportivo

0.45

0.44

 universitaria

0.43

0.42

 indipendente

0.40

 statunit

0.40

//

0.40

 umana

0.40

 fotografico

0.40

Activations Density 0.163%

attends to the token "and" from any preceding token pattern

No Comments

No Known Activations