INDEX

Explanations

references to authority and control structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

SequentialGroup

-0.68

Autoritní

-0.60

featureID

-0.51

#+#

-0.48

acity

-0.47

Искәрмәләр

-0.46

ImageContext

-0.46



-0.45

 autorytatywna

-0.42

 PeEnEo

-0.42

POSITIVE LOGITS

 NSCoder

0.44

 generously

0.44

 kindly

0.42

 respect

0.39

 deliberately

0.36

 opinion

0.36

 generous

0.36

 graciously

0.35

 favor

0.34

FromArgb

0.34

Activations Density 0.095%

references to authority and control structures

No Comments

No Known Activations