INDEX

Explanations

references to hierarchical relationships or structures in data

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

edge label={node[midway,left]

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

boneca

-0.50

awtextra

-0.47

AnchorStyles

-0.47

TargetException

-0.47

WebServlet

-0.46

 ujednoznacz

-0.46

 AssemblyCulture

-0.43

 AssemblyProduct

-0.43

parsedMessage

-0.43

 Exteriores

-0.42

POSITIVE LOGITS

Geplaatst

0.47

WritableDatabase

0.43

 سكانية

0.39

 للاسماء

0.39

casian

0.38

 venuto

0.36

 సౌకర్య

0.36

vician

0.36

 Winaray

0.35

Yin

0.35

Activations Density 0.001%

references to hierarchical relationships or structures in data

edge label={node[midway,left]

No Comments

No Known Activations