INDEX

Explanations

phrases related to recommendations or suggestions about images

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

Central Java

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 Malaysia

-0.41

 Malaysian

-0.34

Malaysia

-0.34

 impos

-0.33

Sdn

-0.32

 Singapore

-0.32

heb

-0.32

 Philippines

-0.31

MODULE

-0.31

 normal

-0.31

POSITIVE LOGITS

 שוליים

0.73

migrationBuilder

0.62

styleType

0.62

UnsafeEnabled

0.59

 Monfieur

0.59

BufferException

0.58

adpleegd

0.57



0.57

 avoient

0.57



0.57

Activations Density 0.001%

phrases related to recommendations or suggestions about images

Central Java

No Comments

No Known Activations