INDEX

Explanations

elements related to specific coding or programming constructs

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

following words

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

principalColumn

-0.64

 FetchType

-0.62

 الرياضيه

-0.56

horabuena

-0.54

 BoxFit

-0.53

parsedMessage

-0.52

 виправивши

-0.52

 EClass

-0.50

 insuffisamment

-0.49

ngrx

-0.48

POSITIVE LOGITS

<tr>

1.19

 surla

0.56

<table>

0.52

GOTREF

0.50

hline

0.48

MathML

0.48

array

0.47

beren

0.46

 gutter

0.45

 Carcass

0.44

Activations Density 0.003%

elements related to specific coding or programming constructs

following words

No Comments

No Known Activations