INDEX

Explanations

the word "for" and variations of it, indicating references to conditions or purposes

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

phrases beginning with for

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

parsedMessage

-0.64

Personensuche

-0.64



-0.62

ValueStyle

-0.59

 ujednoznacz

-0.59

 endregion

-0.57

 Wikimedijinoj

-0.55

irot

-0.55

hombo

-0.55

 HasFactory

-0.54

POSITIVE LOGITS

For

0.98

For

0.93

FOR

0.72

for

0.69

Για

0.63

FOR

0.62

Für

0.62

 Für

0.59

 Για

0.56

for

0.55

Activations Density 0.067%

the word "for" and variations of it, indicating references to conditions or purposes

phrases beginning with for

No Comments

No Known Activations