INDEX

Explanations

instances of the word "allowed."

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 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

complexContent

-0.57

upsi

-0.51

mtrl

-0.50

 akka

-0.50

 MonoBehaviour

-0.50

mobileqq

-0.47

WireFormatLite

-0.47

BASEPATH

-0.46

otesi

-0.46

picasso

-0.46

POSITIVE LOGITS

allowed

0.77

 Allowed

0.72

 ALLOWED

0.69

Allowed

0.69

 allowed

0.68

ALLOWED

0.53

Expected

0.52

 Expected

0.48

nö

0.47

 Acceptable

0.47

Activations Density 0.091%

instances of the word "allowed."

No Comments

No Known Activations