INDEX

Explanations

attends to tokens indicating concepts or actions from related tokens that signify relationships or states

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.41

1:0.11

2:0.08

3:0.07

4:0.09

5:0.03

6:0.07

7:0.11

Negative Logits

Bindable

-0.23

MarshalTo

-0.23

eleste

-0.23

bardier

-0.23

ьаж

-0.22

 soeur

-0.21

EndTag

-0.21

脚注の使い方

-0.21

 veille

-0.21

utete

-0.21

POSITIVE LOGITS

KURZBESCHREIBUNG

0.29

 propOrder

0.28

ThroughAttribute

0.27

HandlerContext

0.26

"}")

0.26

ropoda

0.26

EndGlobalSection

0.25

 Reine

0.25

Архівовано

0.24

 onCancelled

0.24

Activations Density 1.366%

attends to tokens indicating concepts or actions from related tokens that signify relationships or states

No Comments

No Known Activations