INDEX

Explanations

intensifiers that emphasize quality or degree

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

incredibly followed by adjectives

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

 parlor

-0.52

hesis

-0.48

Juice

-0.48

 domo

-0.47

fax

-0.47

Stem

-0.46

Hose

-0.46

View

-0.46

Suit

-0.45

Use

-0.45

POSITIVE LOGITS

 incredibly

1.52

dibly

1.03

 extraordinarily

1.00

 enormously

0.96

 unbelievably

0.94

 extremely

0.91

 unglaub

0.89

extremely

0.88

 amazingly

0.88

 exceedingly

0.88

Activations Density 0.004%

intensifiers that emphasize quality or degree

incredibly followed by adjectives

No Comments

No Known Activations