INDEX

Explanations

attends to information related to confidentiality from tokens that provide context or details about individuals

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

0:0.11

1:0.15

2:0.12

3:0.08

4:0.09

5:0.03

6:0.10

7:0.28

Negative Logits

 فريبيس

-0.32

reonine

-0.30

prefixer

-0.28

 jLabel

-0.27

\}_{

-0.27

ServletException

-0.27

XmlAccessorType

-0.27

TextBoxColumn

-0.27

enoord

-0.27

ensement

-0.26

POSITIVE LOGITS

SharedDtor

0.23

 professionali

0.23

 muñeco

0.23

 volontaire

0.22

 Alamy

0.21

 occhiali

0.21

TagMode

0.21

 tangga

0.21

 Jurí

0.21

 puol

0.21

Activations Density 0.308%

attends to information related to confidentiality from tokens that provide context or details about individuals

No Comments

No Known Activations