INDEX

Explanations

references to arXiv and assembly-related information

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

"arXiv" followed by a publication identifier

np_acts-logits-general · gemini-2.0-flash

references to arXiv preprints

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

getOutputStream

-0.47

 aDecoder

-0.46

 much

-0.44

 Land

-0.42

HostException

-0.41

>{@

-0.41

NSYLVANIA

-0.41

takers

-0.41



-0.40

POSITIVE LOGITS

arXiv

1.96

 CreateTagHelper

1.08

 EconPapers

0.89

 bezeichneter

0.86

 estekak

0.82

 مرئيه

0.82

 оригіналу

0.74

Geplaatst

0.74

 فريبيس

0.73

 الاطلاع

0.73

Activations Density 0.060%

references to arXiv and assembly-related information

"arXiv" followed by a publication identifier

references to arXiv preprints

No Comments

No Known Activations