INDEX

Explanations

references to spouses, particularly the word "wife."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

 please

-0.51

IsMutable

-0.50

 remember

-0.48

 pleased

-0.47

 <<<<<<<<<<<<<<

-0.45

 PLEASE

-0.45

 hurry

-0.44

Ple

-0.44

 հղումներ

-0.43

please

-0.42

POSITIVE LOGITS

 wife

1.52

wife

1.27

Wife

1.25

 Wife

1.23

 WIFE

1.10

 wives

1.03

 esposa

0.96

 moglie

0.94

妻

0.86

妻子

0.84

Activations Density 0.096%

references to spouses, particularly the word "wife."

No Comments

No Known Activations