INDEX

Explanations

specific numerical data and references related to location and institution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.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

urlopen

-0.49

ņem

-0.49

äure

-0.46

 technics

-0.45

 alak

-0.45

ickle

-0.44

 vertra

-0.44

aság

-0.44

 offence

-0.43

Kapitel

-0.43

POSITIVE LOGITS

ंदीखरीदारी

0.68

 MotionEvent

0.68

Datuak

0.62

 المعيارى

0.61

 esternos

0.61

 CreateTagHelper

0.58

الحياه

0.56

IUrlHelper

0.56

niająca

0.56

findpost

0.55

Activations Density 0.009%

specific numerical data and references related to location and institution

No Comments

No Known Activations