INDEX

Explanations

specific nouns and terms associated with unique entities or categories

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

Christmas decorations

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

位

-0.27

 verv

-0.26

 ultimo

-0.26

trusted

-0.26

trust

-0.26

 Counting

-0.26

 ExecuteAsync

-0.26

 szóci

-0.25

 gráfico

-0.25

transition

-0.25

POSITIVE LOGITS

 OMITBAD

0.65

iſen

0.63

contentLoaded

0.63

ثيق

0.61

锈钢

0.59

]=>

0.59

ніципалі

0.58

 kasarigan

0.57

EndProject

0.57

 Normdatei

0.57

Activations Density 0.201%

specific nouns and terms associated with unique entities or categories

Christmas decorations

No Comments

No Known Activations