INDEX

Explanations

attends to tokens with the substring "der" from tokens that follow

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

0:0.71

1:0.01

2:0.01

3:0.01

4:0.11

5:0.06

6:0.02

7:0.02

Negative Logits

AndEndTag

-0.45

includegraphics

-0.42

 HasFactory

-0.42

substack

-0.41

ChildScrollView

-0.40

 latest

-0.40

 arşivlendi

-0.39

XmlAccessType

-0.39

↩

-0.38

featureID

-0.38

POSITIVE LOGITS

<eos>

0.40

tagez

0.31

 almo

0.29

laient

0.27

 also

0.26

ếng

0.26

bezpiecz

0.25

the

0.25

ppure

0.25

Activations Density 0.001%

attends to tokens with the substring "der" from tokens that follow

No Comments

No Known Activations