INDEX

Explanations

attends to the token "split" from various surrounding contexts, often related to division or separation in different scenarios

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61

Prompts (Dashboard)

16,384 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

0:0.01

1:0.01

2:0.55

3:0.02

4:0.00

5:0.00

6:0.08

7:0.04

8:0.00

9:0.01

10:0.00

11:0.00

12:0.00

13:0.10

14:0.08

15:0.00

Negative Logits

out

-0.28

Zel

-0.26

 sides

-0.25

Rob

-0.25

up

-0.24

in

-0.24

不多

-0.24

two

-0.24

 into

-0.24

to

-0.24

POSITIVE LOGITS

webElementXpaths

0.27

verwijspagina

0.26

EndContext

0.24

 ExecuteAsync

0.24

ReusableCell

0.23

 bezeichneter

0.23

usermodel

0.23

]=>

0.21

itemBuilder

0.21

@"/

0.21

Activations Density 0.016%

attends to the token "split" from various surrounding contexts, often related to division or separation in different scenarios

No Comments

No Known Activations