INDEX

Explanations

programming-related terms and structures, especially those involving conditions and null checks

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

checking for existence

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 للاسماء

-0.61

rungsseite

-0.61

complexContent

-0.57

NameInMap

-0.57

RectangleBorder

-0.55

Gambas

-0.55

ſcher

-0.53

niſſe

-0.51

paign

-0.50

raisemb

-0.50

POSITIVE LOGITS

 دیکھیے

0.38

رشف

0.36

EndTag

0.35

InstrumentedTest

0.35

 typelib

0.35

AccessorTable

0.34

TAL

0.31

 while

0.31

 ménages

0.30

ImageContext

0.29

Activations Density 0.017%

programming-related terms and structures, especially those involving conditions and null checks

checking for existence

No Comments

No Known Activations