INDEX

Explanations

attends to closing double asterisks from various opening tokens, indicating a relation to code or structure points in the text

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

0:0.31

1:0.25

2:0.07

3:0.08

4:0.06

5:0.06

6:0.06

7:0.07

Negative Logits

Personendaten

-0.38

$_['

-0.35

 endblock

-0.32

oredCriteria

-0.29

DockStyle

-0.28

Slf

-0.28

wußt

-0.28

setupUi

-0.28

elemField

-0.27

WriteTagHelper

-0.27

POSITIVE LOGITS

tagHelperRunner

0.36

 initComponents

0.35

BnF

0.33

tition

0.30

rethe

0.29

 getAmount

0.29

sigen

0.29

 surla

0.29

mment

0.28

Activations Density 0.400%

attends to closing double asterisks from various opening tokens, indicating a relation to code or structure points in the text

No Comments

No Known Activations