INDEX

Explanations

attends to markers from new or minor tokens regardless of order in the sentence

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

0:0.13

1:0.15

2:0.05

3:0.07

4:0.09

5:0.29

6:0.06

7:0.12

Negative Logits

ed

-0.28

AutoSize

-0.27

 bomb

-0.26

<h4>

-0.26

FI

-0.25

정

-0.25

 Dumnezeu

-0.24

CCS

-0.24

 GenerationType

-0.24

-0.23

POSITIVE LOGITS

rungsseite

0.49

sidemargin

0.48

Geplaatst

0.47

VIAF

0.46

 <=",

0.44

Personensuche

0.43

liefert

0.43

 Roskov

0.43

 hObject

0.42

 protoimpl

0.41

Activations Density 0.000%

attends to markers from new or minor tokens regardless of order in the sentence

No Comments

No Known Activations