INDEX

Explanations

references to app availability and user experiences

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

download app

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

EncBinary

-0.45

GIVEREF

-0.40

hamcrest

-0.38

GEBURTSDATUM

-0.36

 stretching

-0.35

 typedef

-0.34

fromDate

-0.34

 tratados

-0.34

 stretched

-0.34

 Ayres

-0.34

POSITIVE LOGITS

 AssemblyVersion

0.58

StoryboardSegue

0.56

UserScript

0.55

RegressionTest

0.55

bootstrapcdn

0.54

السكان

0.50

RTSN

0.50

AppMethodBeat

0.49

 queſta

0.48

cerely

0.48

Activations Density 0.035%

references to app availability and user experiences

download app

No Comments

No Known Activations