INDEX

Explanations

the word "take" in various contexts, indicating a focus on actions or requests

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dhism

-0.82

 koszulka

-0.79

theless

-0.78

timis

-0.77

etheless

-0.77

PMS

-0.77

 Sams

-0.77

óval

-0.76

PMS

-0.75

 Horner

-0.73

POSITIVE LOGITS

take

2.48

 take

2.40

Take

2.39

Taking

2.33

 Take

2.29

taking

2.24

 TAKE

2.23

 Taking

2.22

 taking

2.21

TAKE

2.14

Activations Density 0.109%

the word "take" in various contexts, indicating a focus on actions or requests

No Comments

No Known Activations

the word "take" in various contexts, indicating a focus on actions or requests

No Comments

No Known Activations