INDEX

Explanations

attends from property-related tokens to corresponding identifier tokens

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

0:0.12

1:0.17

2:0.13

3:0.10

4:0.08

5:0.03

6:0.07

7:0.26

Negative Logits

 AssemblyCulture

-0.42

 PeEnEo

-0.38

 Paglinawan

-0.38

elemField

-0.37

LookAnd

-0.37

"):

-0.36

Datuak

-0.35

']))

-0.35

Carriera

-0.34

 lenker

-0.34

POSITIVE LOGITS

OrderDetail

0.31

 onderdeel

0.27

vese

0.26

continu

0.25

lacian

0.24

 zwar

0.24

sld

0.24

 Broadway

0.23

Beschreibung

0.23

 excepting

0.23

Activations Density 0.096%

attends from property-related tokens to corresponding identifier tokens

No Comments

No Known Activations