INDEX

Explanations

non-zeros indicating frequency or emphasis in text

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_7/width_16k/average_l0_92

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

AddTagHelper

-0.53

 darbu

-0.48

 للمعارف

-0.48

HasForeignKey

-0.48

 miniaturka

-0.48

foglal

-0.47

parsedMessage

-0.46

évaluateur

-0.46

 kapture

-0.46

 disponibilità

-0.46

POSITIVE LOGITS

 fiery

0.45

 brief

0.45

ChrTalk

0.44

 eerie

0.43

 soulful

0.43

athery

0.42

 Bourgeois

0.42

non

0.42

 idealistic

0.42

 acoustic

0.41

Activations Density 0.168%

non-zeros indicating frequency or emphasis in text

No Known Activations