INDEX

Explanations

references to architectural features and openings in a structure

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

\{\\

-0.74

tagHelperRunner

-0.73

ValueStyle

-0.69

 CURIAM

-0.66

+#+

-0.66

cshtml

-0.61

UnknownFieldSet

-0.60

oa̍t

-0.60

Trama

-0.59

amaged

-0.59

POSITIVE LOGITS

 directly

0.60

 bezpośrednio

0.54

 direct

0.49

 kasarigan

0.49

directly

0.46

 مستقیم

0.46

 مباشرة

0.44

 ViewPager

0.44

sizeCache

0.43

 directamente

0.43

Activations Density 0.013%

references to architectural features and openings in a structure

No Comments

No Known Activations