INDEX

Explanations

attends to newline characters from various tokens that represent code or format elements

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.36

1:0.02

2:0.02

3:0.02

4:0.16

5:0.09

6:0.05

7:0.23

Negative Logits

 kaarangay

-0.45

RegressionTest

-0.42

 pinulongan

-0.42

aarrggbb

-0.42

 سكانية

-0.42

skosten

-0.40

quelize

-0.40

WebVitals

-0.39

RetentionPolicy

-0.39

WriteTagHelper

-0.38

POSITIVE LOGITS

 ProtoMessage

0.31

<strong>

0.30

Gön

0.30

BeginInit

0.30

 Wikimedijinoj

0.30

 atonement

0.26

 Engl

0.26

 nitrous

0.26

<u>

0.26

Hochspringen

0.25

Activations Density 0.023%

attends to newline characters from various tokens that represent code or format elements

No Comments

No Known Activations