INDEX

Explanations

scheduling and timing details

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_18/width_65k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.18.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

 menstrual

-0.91

 Moon

-0.89

 lunar

-0.88

 menstruation

-0.86

 menstru

-0.85

 moon

-0.85

Moon

-0.82

moon

-0.77

 MOON

-0.75

 cycle

-0.75

POSITIVE LOGITS

 typelib

0.75

findpost

0.68

XmlAccessType

0.67

endphp

0.67

protoc

0.67

 Infórmanos

0.66

BeginContext

0.62

WriteTagHelper

0.61

帖最后由

0.60

ngdoc

0.59

Activations Density 0.132%

scheduling and timing details

No Comments

No Known Activations