INDEX

Explanations

attends to boolean-related tokens from general function-related tokens

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.03

1:0.04

2:0.07

3:0.06

4:0.15

5:0.55

6:0.03

7:0.03

Negative Logits

 myſelf

-1.35

 itſelf

-1.34

 ―――――

-1.26

Efq

-1.25

 faſt

-1.22

 ་་

-1.19

 ſind

-1.19

^(@)

-1.19

 Monfieur

-1.18

 Theſe

-1.17

POSITIVE LOGITS

0.80

0.79

↵

0.79

and

0.76

the

0.76

↵↵

0.75

0.72

in

0.67

0.66

Activations Density 0.055%

attends to boolean-related tokens from general function-related tokens

No Comments

No Known Activations