INDEX

Explanations

negative numerical values in the context of scores or statistics

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

WebServlet

-0.60

 initComponents

-0.59

ViewFeatures

-0.57

osoba

-0.56

 Vorm

-0.55

ziehen

-0.55

 AssemblyVersion

-0.54

FileVersion

-0.53

 Werk

-0.52

InitStruct

-0.51

POSITIVE LOGITS

 XNUMX

0.66

enumi

0.63

ंदीखरीदारी

0.63

MLLoader

0.62

giveness

0.61

 Carboniferous

0.60

 désolés

0.60

áfrica

0.58

 تضيفلها

0.58

".

0.58

Activations Density 0.007%

negative numerical values in the context of scores or statistics

No Comments

No Known Activations