INDEX

Explanations

attends to terms related to a concept from other terms also related to that concept

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

0:0.08

1:0.10

2:0.11

3:0.10

4:0.10

5:0.12

6:0.10

7:0.24

Negative Logits

DockStyle

-0.36

ActionCreators

-0.32

="?

-0.31

VideoCapture

-0.30

endphp

-0.29

addPreferredGap

-0.28

bleau

-0.27

 سكانية

-0.27

minton

-0.26

bootstrapcdn

-0.26

POSITIVE LOGITS

VIAF

0.30

Huy

0.27

 nahilalakip

0.27

fjspx

0.26

:✨

0.25

 HasFactory

0.24

Viitteet

0.24

 vänt

0.24

 durs

0.23

featureID

0.23

Activations Density 0.000%

attends to terms related to a concept from other terms also related to that concept

No Comments

No Known Activations