INDEX

Explanations

references to specific weeks and frequency of occurrences related to time

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

 Glaser

-0.36

 Gilmour

-0.35

paddingRight

-0.35

 into

-0.33

 Southern

-0.32

dsp

-0.32

 Gentry

-0.31

 Carlson

-0.31

 Dougherty

-0.31

 Tsai

-0.31

POSITIVE LOGITS

Week

1.59

 week

1.46

week

1.41

 Week

1.41

 WEEK

1.39

WEEK

1.39

weeks

1.35

 weeks

1.27

 semana

1.25

 semaine

1.20

Activations Density 0.041%

references to specific weeks and frequency of occurrences related to time

No Comments

No Known Activations