INDEX

Explanations

phrases indicating individuality or uniqueness in various contexts

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

each unique, own, or different

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

PDA

-0.39

ंदीखरीदारी

-0.37

PDA

-0.37

Kjelder

-0.35

Dtor

-0.35

zenten

-0.34

<tfoot>

-0.33

ginate

-0.33

Trend

-0.33

 meeste

-0.33

POSITIVE LOGITS

 defaultstate

0.73

ghijklmnop

0.54

 nonUne

0.52

不同的

0.50

featureID

0.49

GHIJKLM

0.49

 оригіналу

0.49

different

0.48

 Egger

0.47

oncesto

0.47

Activations Density 0.027%

phrases indicating individuality or uniqueness in various contexts

each unique, own, or different

No Comments

No Known Activations