INDEX

Explanations

attends to both-related tokens from both-related tokens

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

0:0.04

1:0.10

2:0.05

3:0.02

4:0.14

5:0.57

6:0.02

7:0.03

Negative Logits

Personensuche

-0.34

Datuak

-0.33

setContentView

-0.30

 MainAxisSize

-0.29

adpleegd

-0.29

 }}"></

-0.29

parsedMessage

-0.28

Chham

-0.27

 raiſ

-0.27



-0.27

POSITIVE LOGITS

 energética

0.19

クリル

0.14

ڑ

0.14

 strona

0.14

 térmico

0.13

 coger

0.13

また

0.13

 líquida

0.13

 encore

0.13

 scuola

0.13

Activations Density 0.678%

attends to both-related tokens from both-related tokens

No Comments

No Known Activations