INDEX

Explanations

the word "whose" and variations indicating possession or belonging

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_2/width_16k/average_l0_141

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

Datuak

-0.68

It

-0.59

it

-0.57

-0.56

kannya

-0.53

 will

-0.53

raltar

-0.52

avadoc

-0.52

-0.51

POSITIVE LOGITS

 whose

2.51

 Whose

2.40

whose

2.40

Whose

2.29

 whofe

2.26

 whoſe

2.11

 cuya

1.88

 cuyas

1.85

 cuyos

1.82

 cuyo

1.81

Activations Density 0.043%

the word "whose" and variations indicating possession or belonging

No Comments

No Known Activations