INDEX

Explanations

references to places of residence or dwelling

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

dwelling and user input

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

verständlich

-0.37

fertigt

-0.36

RegExp

-0.36

Contactez

-0.35

 ferner

-0.34

DecimalFormat

-0.34

Égypte

-0.34

Egypte

-0.34

androidx

-0.34

 estran

-0.34

POSITIVE LOGITS

########.

0.68

 home

0.62

MessageTagHelper

0.61

tagHelperRunner

0.57

 Chwiliwch

0.56

Aholisi

0.56

parsedMessage

0.56

AndEndTag

0.56

GEBURTS

0.55

RegressionTest

0.54

Activations Density 0.001%

references to places of residence or dwelling

dwelling and user input

No Comments

No Known Activations