INDEX

Explanations

phrases indicating achievement or success

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

past / previous / ago

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

󠁳

-0.34

DockStyle

-0.33

AutoresizingMask

-0.32

$_['

-0.31

Ato

-0.31

ThroughAttribute

-0.30

AG

-0.30

➯

-0.30

ча

-0.29

POSITIVE LOGITS

vician

0.55

 kasarigan

0.54

 beroep

0.53

+#+#

0.49

AnimationsModule

0.49

GraphicsUnit

0.48

scalatest

0.46

 Erziehung

0.46

 mijne

0.46

LLocation

0.45

Activations Density 0.023%

phrases indicating achievement or success

past / previous / ago

No Comments

No Known Activations