INDEX

Explanations

specific symbols, numbers, and formatting often related to data representation or statistical information

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

numbers and abbreviations

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

 ComVisible

-0.73

ніципалі

-0.57

bootstrapcdn

-0.57

تقاوى

-0.56

 Meksiku

-0.56

Clik

-0.55

Демографія

-0.55

__':

-0.54

@[+][

-0.54

ApiModelProperty

-0.52

POSITIVE LOGITS

 otomatig

0.40

 gärna

0.39

 fér

0.35

 Audio

0.34

 nada

0.33

 preghiera

0.32

tidig

0.32

σθαι

0.32

 Last

0.31

 Character

0.31

Activations Density 0.018%

specific symbols, numbers, and formatting often related to data representation or statistical information

numbers and abbreviations

No Comments

No Known Activations