INDEX

Explanations

words or phrases indicating personal choices and emotional expressions

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 onto

-0.51

 vēl

-0.50

 interactions

-0.46

 interact

-0.45

 throughout

-0.44

 portions

-0.44

 prior

-0.42

 utilizing

-0.42

 jornadas

-0.41

 pertain

-0.41

POSITIVE LOGITS

oa̍t

0.64

GEBURTSDATUM

0.58

TokenNameLBRACE

0.52

mobileqq

0.52

 tartalomajánló

0.51

GEBURTS

0.49

richTextPanel

0.46

 تضيفلها

0.46

localctx

0.46

Personendaten

0.45

Activations Density 0.009%

words or phrases indicating personal choices and emotional expressions

No Comments

No Known Activations