INDEX

Explanations

phrases related to the beginning or initiation of events

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

start or open

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 Suiza

-0.35

 Diciembre

-0.31

ąt

-0.27

 najdete

-0.27

 Belgien

-0.26

 complémentaires

-0.26

 Opiniones

-0.26

 porcentaje

-0.26

({

-0.26

POSITIVE LOGITS

parsedMessage

0.84

enterOuterAlt

0.81

 initiating

0.75

 StringTokenizer

0.73

offsetof

0.73

 Chwiliwch

0.73

 lenker

0.72

featureID

0.72

 PeEnEo

0.71

ContentLoaded

0.71

Activations Density 0.010%

phrases related to the beginning or initiation of events

start or open

No Comments

No Known Activations