INDEX

Explanations

references to historical events or names of people involved in significant occurrences

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

reporting dates and costs

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

 ब्रेकडाउन

-0.56

دانشنامهٔ

-0.55

 المعيارى

-0.54

 prioriti

-0.53

 Chwiliwch

-0.50

 prioritized

-0.50

 ERSITY

-0.48

balleur

-0.48

 prioritize

-0.48

 contextes

-0.47

POSITIVE LOGITS

CodeAttribute

0.44

addCriterion

0.39

IGENCE

0.36

 aDecoder

0.35

newspaper

0.33

DropColumn

0.32

 iconLine

0.32

 interviewed

0.32

 føl

0.32

TemporalType

0.32

Activations Density 0.022%

references to historical events or names of people involved in significant occurrences

reporting dates and costs

No Comments

No Known Activations