INDEX

Explanations

attends to the token "segment" from various types of tokens indicating classification or division

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

0:0.65

1:0.01

2:0.01

3:0.04

4:0.16

5:0.03

6:0.02

7:0.05

Negative Logits

fuls

-0.32

☸

-0.29

 Перейти

-0.29

full

-0.28

Level

-0.27

λέον

-0.27

zeb

-0.26

Cited

-0.26

 Preferencias

-0.26

 vidi

-0.26

POSITIVE LOGITS

InSection

0.50

 виправивши

0.47

GOTREF

0.43

MemoryWarning

0.41

glBindTexture

0.40

 دیکھیے

0.39

gonic

0.38

 cherchés

0.37

 فريبيس

0.36

]="

0.36

Activations Density 0.006%

attends to the token "segment" from various types of tokens indicating classification or division

No Comments

No Known Activations