INDEX

Explanations

vocabulary related to dominance or power dynamics

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

 escar

-0.44

testdata

-0.40

Legs

-0.39

ientí

-0.38

rdata

-0.38

̯

-0.38

-0.37

Grat

-0.37

 scrap

-0.37

 Alger

-0.36

POSITIVE LOGITS

 dominance

1.18

 dominate

1.06

 dominant

1.05

 dominated

1.02

 domination

1.02

 dominating

1.00

dominant

0.97

 dominates

0.96

domin

0.96

 dominante

0.96

Activations Density 0.254%

vocabulary related to dominance or power dynamics

No Comments

No Known Activations