INDEX

Explanations

attends to function call-related tokens from list operation 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

0:0.13

1:0.15

2:0.39

3:0.04

4:0.06

5:0.05

6:0.03

7:0.10

Negative Logits

varlak

-0.30

toBeDefined

-0.30

젝

-0.29

ftagPool

-0.29

дравствуйте

-0.28

لينكات

-0.28

<thead>

-0.27

 CreateTagHelper

-0.27

 Hamp

-0.27

stacles

-0.27

POSITIVE LOGITS

 surla

0.33

ValueStyle

0.30

 Wicidata

0.30

Personendaten

0.29

 transfieras

0.28

MessageState

0.28

halb

0.26

 refusé

0.26

 catedral

0.25

awtextra

0.25

Activations Density 0.086%

attends to function call-related tokens from list operation tokens

No Comments

No Known Activations