INDEX

Explanations

references to programming concepts or constructs

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

 AssemblyVersion

-0.44

 forState

-0.36

batore

-0.35

 propOrder

-0.34

CloseOperation

-0.34

 <=",

-0.34

 Forscher

-0.33

หลวง

-0.33

extAlignment

-0.33

 unpublished

-0.33

POSITIVE LOGITS

OGND

0.59

IntoConstraints

0.56

 jspb

0.48

NameInMap

0.46

styleType

0.46

ãng

0.46

fycat

0.45

SizeF

0.45

 epis

0.44

 NUKAT

0.44

Activations Density 0.001%

references to programming concepts or constructs

No Comments

No Known Activations