INDEX

Explanations

references to scientific nomenclature or classification

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

user input markers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 demais

-0.38

 potest

-0.36

 habet

-0.34

paisaje

-0.34

 Lijst

-0.32

Билгалдахарш

-0.32

 Kraw

-0.30

 cuesta

-0.30

 finishing

-0.30

Produzione

-0.30

POSITIVE LOGITS

errHandler

0.60

HtmlAttribute

0.58

AndEndTag

0.57

 LoginForm

0.56

Gön

0.55

 kasarigan

0.54

setPointSize

0.53

DockStyle

0.52

maids

0.52

 Elegance

0.52

Activations Density 0.020%

references to scientific nomenclature or classification

user input markers

No Comments

No Known Activations