INDEX

Explanations

instances of conjunctions and punctuation indicative of lists or separations

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

list descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

GEBURTSDATUM

-0.40

⮕

-0.39

OGND

-0.39

Portale

-0.38

 volume

-0.38

 Infórmanos

-0.38

volume

-0.37

featureID

-0.37

 cobre

-0.36

 beweging

-0.36

POSITIVE LOGITS

 Corbu

0.48

 Ebenso

0.46

berdayakan

0.45

ConverterFactory

0.44

bonucle

0.44

tencent

0.43

BindView

0.43

ImageField

0.43

oprecip

0.42

rily

0.42

Activations Density 0.086%

instances of conjunctions and punctuation indicative of lists or separations

list descriptions

No Comments

No Known Activations