INDEX

Explanations

questions about identity and role

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

 Stedman

-0.44

 scant

-0.40

 geçen

-0.39

 preventivo

-0.38

只是一

-0.38

phazard

-0.38

 passagem

-0.37

 massive

-0.37

다는

-0.36

aptation

-0.36

POSITIVE LOGITS

Who

0.86

rungsseite

0.85

Who

0.84

who

0.83

who

0.77

 hvem

0.76

Quién

0.76

Quiénes

0.75

 whom

0.74

 whome

0.73

Activations Density 0.065%

questions about identity and role

No Comments

No Known Activations