INDEX

Explanations

various numeric patterns and formatting, especially in relation to addresses and numerical values

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

floor numbers and addresses

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

 tartalomajánló

-0.61

脚注の使い方

-0.61

setVerticalGroup

-0.59

OGND

-0.59

AutoScaleMode

-0.57

TagMode

-0.57

يميديا

-0.54

fromnode

-0.54

TypedDataSet

-0.53

POSITIVE LOGITS

 apartments

0.35

 apartment

0.32

 excuse

0.30

eX

0.30

Apartment

0.30

down

0.29

Down

0.29

FBS

0.29

 yours

0.28

 Apartment

0.28

Activations Density 0.006%

various numeric patterns and formatting, especially in relation to addresses and numerical values

floor numbers and addresses

No Comments

No Known Activations