INDEX

Explanations

phrases related to revealing or exposing hidden truths or features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 23-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_23/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.23.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

 EconPapers

-0.39

Hentet

-0.35

redited

-0.34

Zusammenfassung

-0.34

Datuak

-0.33

olak

-0.32

ситет

-0.32

 будешь

-0.31

 donuts

-0.31

ariant

-0.31

POSITIVE LOGITS

 underlying

1.05

underlying

0.92

 Underlying

0.83

Underlying

0.82

 underneath

0.78

 behind

0.66

 darunter

0.65

 underlies

0.65

parsedMessage

0.64

behind

0.63

Activations Density 0.712%

phrases related to revealing or exposing hidden truths or features

No Comments

No Known Activations