INDEX

Explanations

instances of being told or asked about experiences and situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

 logra

-0.37

 absolument

-0.37

 envers

-0.37

 ambient

-0.36

 cosméticos

-0.36

Contribu

-0.36

 giro

-0.36

 contribución

-0.36

 aggregate

-0.36

 matin

-0.35

POSITIVE LOGITS

CppMethod

0.57

 diminta

0.54

webElement

0.53

被迫

0.53

RoutingModule

0.52

PerformLayout

0.50

onOptions

0.50

 RemoteException

0.49

 recie

0.47

лтемелер

0.47

Activations Density 0.335%

instances of being told or asked about experiences and situations

No Comments

No Known Activations