INDEX

Explanations

phrases related to legal accusations and investigations

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

constructions that express causation or explanation (using "be" as a linking verb in passive voice patterns).

oai_token-act-pair · claude-4-5-haiku Triggered by @emiglarou

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.31.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

 seemingly

-0.50

 казалось

-0.46

 aparentemente

-0.40

SharedCtor

-0.40

VIO

-0.40

 dinilai

-0.39

 للاس

-0.39

 przekon

-0.39

NameValuePair

-0.39

 wydaje

-0.39

POSITIVE LOGITS

 belonged

0.68

 belong

0.62

Tembelea

0.61

 fraude

0.60

 apparten

0.59

 fake

0.58

 faked

0.57

 genuine

0.57

 belonging

0.57

 pertence

0.55

Activations Density 0.881%

phrases related to legal accusations and investigations

constructions that express causation or explanation (using "be" as a linking verb in passive voice patterns).

No Comments

No Known Activations