INDEX

Explanations

programming or code structure elements

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 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

ValueStyle

-0.66

oa̍t

-0.57

Diweddarwch

-0.51

 AssemblyCompany

-0.51

 الاطلاع

-0.51

XtraGrid

-0.49

 iNdEx

-0.49



-0.48

 محفوظة

-0.48

ChildScrollView

-0.48

POSITIVE LOGITS

.*")]

0.58

 FetchType

0.53

InstrumentedTest

0.52

참고

0.46

 HasFactory

0.46

GenerationType

0.43

 Савезне

0.43

fromnode

0.41

zato

0.39

istical

0.37

Activations Density 0.280%

programming or code structure elements

No Comments

No Known Activations