INDEX

Explanations

punctuation marks, specifically periods

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

user input delimiters

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

]--;

-0.45

 praça

-0.38

 gouver

-0.38

WindowListener

-0.36

GenerationType

-0.36

']));

-0.36

initializeApp

-0.36

 sauve

-0.35

gridColumn

-0.35

CardBody

-0.35

POSITIVE LOGITS

loyees

0.61

ChildScrollView

0.60

 EconPapers

0.58

 nahilalakip

0.57

liceerd

0.56

Jîn

0.54

localctx

0.52

adaptiveStyles

0.51

Carriera

0.51

出版年

0.50

Activations Density 0.000%

punctuation marks, specifically periods

user input delimiters

No Comments

No Known Activations