INDEX

Explanations

high activation sections of text, likely reflecting key concepts or critical points

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 BoxFit

-0.79

findpost

-0.71

SourceChecksum

-0.69

HostException

-0.62

 Numerade

-0.61

isContained

-0.61

 DbSet

-0.60

الإنجليزية

-0.59

GTCX

-0.59

олові

-0.58

POSITIVE LOGITS

 mères

0.39

Dial

0.33

 pères

0.33

 ejemplar

0.32

 miej

0.31

 genética

0.31

Czytaj

0.31

Season

0.31

 âmes

0.31

 enfans

0.30

Activations Density 0.000%

high activation sections of text, likely reflecting key concepts or critical points

No Comments

No Known Activations