INDEX

Explanations

sequence markers and transitional phrases in a narrative context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 purpoſe

-0.95

 juſt

-0.93

 reaſon

-0.88

ſelf

-0.88

 whoſe

-0.84

 himſelf

-0.83

 Monfieur

-0.80

 becauſe

-0.79

 cauſe

-0.79

 itſelf

-0.78

POSITIVE LOGITS

 then

1.37

 Then

1.32

 THEN

1.23

Then

1.20

 puis

1.17

 followed

1.15

 Followed

1.10

Followed

1.10

followed

1.09

 ثم

1.07

Activations Density 0.339%

sequence markers and transitional phrases in a narrative context

No Comments

No Known Activations