INDEX

Explanations

words associated with legal or regulatory proceedings and scientific terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

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

ებ

-1.01

RectangleBorder

-1.00

 BoxDecoration

-0.97

PhysRevD

-0.94

 MainAxisSize

-0.92

 архивлан

-0.85

 ProtoMessage

-0.83

Geplaatst

-0.82

 snippetHide

-0.81

بوابة

-0.79

POSITIVE LOGITS

 cadeira

0.55

0.53

 ordres

0.53

 Stead

0.53

pleen

0.52

円

0.51

 McLeod

0.51

 ouvrages

0.49

 prêtres

0.49

Activations Density 0.013%

words associated with legal or regulatory proceedings and scientific terminology

No Comments

No Known Activations