INDEX

Explanations

references to average values or averages in various contexts

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

 average

-1.77

average

-1.13

theless

-0.88

 promedio

-0.84

.)}

-0.81

 moyenne

-0.80

Hochspringen

-0.74

 durchschnitt

-0.73

 only

-0.71

ίναι

-0.71

POSITIVE LOGITS

 arroz

0.65

legacy

0.62

 desaparecido

0.60

 testamento

0.60

 comprando

0.59

padu

0.59

bebasan

0.59

auteuil

0.59

 guardado

0.58

 cuid

0.58

Activations Density 0.024%

references to average values or averages in various contexts

No Comments

No Known Activations