INDEX

Explanations

references to studies or evidence in academic writing

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

citations

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

írus

-0.34

 собі

-0.34

dor

-0.31

 gérer

-0.31

 self

-0.30

endpush

-0.30

ehr

-0.30

WL

-0.30

xture

-0.30

addContainerGap

-0.29

POSITIVE LOGITS

참고

0.67

 Савезне

0.66

 resourceCulture

0.65

MessageOf

0.63

 مشين

0.63

 references

0.59

SourceChecksum

0.57

 autorytatywna

0.55

 ModelExpression

0.54

Viitteet

0.54

Activations Density 0.976%

references to studies or evidence in academic writing

citations

No Comments

No Known Activations