INDEX

Explanations

phrases indicating the subject's identity and status

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

was/is followed by role or past state

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

setVerticalGroup

-0.49

uxxxx

-0.48

ніципалі

-0.46

хьтан

-0.43

Bioaccumulative

-0.40

ACHUSET

-0.39

ValueGeneration

-0.39

berkeley

-0.39

DataAnnotations

-0.39

发表于

-0.38

POSITIVE LOGITS

 mukana

0.57

Efq

0.51

 seorang

0.47

 uveden

0.47

 approbation

0.46

 Majefty

0.45

 Administrativna

0.45

 engaged

0.45

 capable

0.45

 able

0.45

Activations Density 0.023%

phrases indicating the subject's identity and status

was/is followed by role or past state

No Comments

No Known Activations