INDEX

Explanations

attends to educational-related tokens from status-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

0:0.46

1:0.02

2:0.07

3:0.02

4:0.21

5:0.14

6:0.01

7:0.03

Negative Logits

 itſelf

-0.71

 myſelf

-0.65

 Jefus

-0.65

 Theſe

-0.62

 Shakspeare

-0.61

$_"

-0.60

 Houſe

-0.60

 Anſ

-0.60

 Monfieur

-0.59

 photolibrary

-0.56

POSITIVE LOGITS

<eos>

0.41

↵

0.40

Ver

0.39

0.37

en

0.36

de

0.35

rep

0.35

Activations Density 0.009%

attends to educational-related tokens from status-related tokens

No Comments

No Known Activations