INDEX

Explanations

phrases related to watching and revisiting movies or performances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ſche

-0.57

 ſta

-0.54

 Reſ

-0.52

ustain

-0.49

 Majefty

-0.48

ſelf

-0.46

 Perſ

-0.46

 Leaflet

-0.46

 Houſe

-0.45

 Saver

-0.45

POSITIVE LOGITS

 reread

0.55

 rewatch

0.46

脚注の使い方

0.41

twimg

0.40

tagext

0.39

 архивлан

0.38

癖

0.38

 repetir

0.38

 repeat

0.37

再來

0.36

Activations Density 0.013%

phrases related to watching and revisiting movies or performances

No Comments

No Known Activations