INDEX

Explanations

references to political arrests and activists

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

accused and jailed

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

 الشر

-0.35

SourceChecksum

-0.34

 فريبيس

-0.34

 العنوان

-0.33

 hata

-0.33

addCriterion

-0.33

awakeFromNib

-0.32

InputTagHelper

-0.32

yntaxException

-0.32

SQLException

-0.31

POSITIVE LOGITS

 imprison

0.51

❋

0.48

accu

0.47

 accused

0.46

Guilty

0.46

 jailed

0.46

servant

0.45

 imprisonment

0.45

Arrest

0.45

Pozdrawiam

0.44

Activations Density 0.097%

references to political arrests and activists

accused and jailed

No Comments

No Known Activations