INDEX

Explanations

numerical data points, specifically in tabular or structured formats

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

references to figures or tables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

".

-0.56

 cern

-0.53

cellona

-0.50

äsche

-0.49

LLocation

-0.49

?";

-0.48

uſ

-0.47

 poffe

-0.47

 custody

-0.46

سطس

-0.46

POSITIVE LOGITS

0.61

RegistryLite

0.60

netbeans

0.60

 Paglinawan

0.59

 mattino

0.56

hasMoreElements

0.53

Referencias

0.52

Παραπομπές

0.52

ilaian

0.50

 المعيارى

0.50

Activations Density 0.121%

numerical data points, specifically in tabular or structured formats

references to figures or tables

No Comments

No Known Activations