INDEX

Explanations

punctuation marks and other characters that indicate structure or notation in text

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

names followed by publication

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

IsUnicode

-0.31

 bateaux

-0.28

 nivå

-0.28

wendungs

-0.27

 navires

-0.27

 enfans

-0.27

 épaules

-0.27

 degrés

-0.26

 requieren

-0.26

 aktivitet

-0.26

POSITIVE LOGITS

 Meksiku

0.71

0.67

DIPSETTING

0.66

:✨

0.66

 ſchon

0.65

styleUrls

0.65

RectangleBorder

0.65

OGND

0.63

ViewFeatures

0.61

 パンチラ

0.61

Activations Density 0.030%

punctuation marks and other characters that indicate structure or notation in text

names followed by publication

No Comments

No Known Activations