INDEX

Explanations

attends to the numerical score from the decision of a match

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

0:0.11

1:0.17

2:0.11

3:0.09

4:0.13

5:0.08

6:0.10

7:0.17

Negative Logits

MessageTagHelper

-0.29

Personendaten

-0.26

 ligiloj

-0.26

right

-0.25

 protoimpl

-0.25

wpi

-0.24

 GenerationType

-0.23

妬

-0.23

homonymie

-0.22

addContainerGap

-0.21

POSITIVE LOGITS

STANDING

0.28

fonos

0.28

Referanser

0.27

 kereszt

0.27

tvguidetime

0.26

Lähteet

0.26

ükemmel

0.26

couverte

0.26

nesia

0.25

czaj

0.25

Activations Density 0.050%

attends to the numerical score from the decision of a match

No Comments

No Known Activations