INDEX

Explanations

names of individuals involved in criminal activities

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

Names followed by verbs or commas

np_acts-logits-general · gemini-2.0-flash

names in legal contexts

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 AssemblyCulture

-0.42

 مرئيه

-0.41

pushFollow

-0.39

 karşılaş

-0.36

boa

-0.36

tomans

-0.36

 maravilloso

-0.36

mover

-0.36

lamy

-0.35

 THERE

-0.35

POSITIVE LOGITS

MethodManager

0.48

 accused

0.48

 deputies

0.47

__(/*!

0.46

 arrest

0.45

 allegedly

0.45

 alleged

0.45

 dilaporkan

0.44

 boyfriend

0.44

Бахар

0.44

Activations Density 0.299%

names of individuals involved in criminal activities

Names followed by verbs or commas

names in legal contexts

No Comments

No Known Activations