INDEX

Explanations

attends to distinct notable features or actions from commonly general instructions

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

0:0.34

1:0.16

2:0.08

3:0.06

4:0.06

5:0.04

6:0.09

7:0.14

Negative Logits

addContainerGap

-0.35

 endnu

-0.33

 cauſe

-0.33

 colectiva

-0.33

 purpoſe

-0.32

pausal

-0.32

 omnia

-0.32

 Shoulders

-0.32

GEBURTSDATUM

-0.32

Carreira

-0.31

POSITIVE LOGITS

 disambiguazione

0.34

RTSC

0.31

Identyfik

0.30

̣

0.30

ritsar

0.30

évaluateur

0.30

 sendiri

0.29

ker

0.29

sizeCache

0.29

="@+

0.28

Activations Density 0.255%

attends to distinct notable features or actions from commonly general instructions

No Comments

No Known Activations