INDEX

Explanations

occurrences of the word "mail" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

SourceChecksum

-0.90

 betweenstory

-0.82

tructuring

-0.81

lernen

-0.78

 Clapton

-0.77

JSONException

-0.77

cym

-0.76

ueux

-0.76

roves

-0.75

 Osbourne

-0.74

POSITIVE LOGITS

 MAIL

1.97

 Mail

1.96

 mail

1.92

mail

1.83

Mail

1.80

MAIL

1.69

 mails

1.47

mails

1.24

Mails

1.24

 mailed

1.14

Activations Density 0.055%

occurrences of the word "mail" and related terms

No Comments

No Known Activations