INDEX

Explanations

interrogative words and phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 Houſe

-0.58

ValueStyle

-0.47

 houſe

-0.47

 raiſ

-0.44

RegressionTest

-0.40

StackNavigator

-0.40

 ſmall

-0.39

 preſent

-0.38

piecz

-0.38

ivil

-0.37

POSITIVE LOGITS

屉

0.59

DockStyle

0.58

 Does

0.57

 does

0.56

 للمعارف

0.56

 Didi

0.56

idemiology

0.55

 need

0.55

do

0.55

did

0.54

Activations Density 0.221%

interrogative words and phrases

No Comments

No Known Activations