INDEX

Explanations

references to previous relationships or past associations with the term "former."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

enumi

-0.47

 Vielfalt

-0.42

jalá

-0.41

 them

-0.40

httphttps

-0.40

 mourut

-0.40

libatkan

-0.39

 typelib

-0.39

 Normdatei

-0.39

Население

-0.39

POSITIVE LOGITS

interopRequire

0.51

UrlResolution

0.50

 svých

0.50

 constexpr

0.49

 remaining

0.49

ActionCreators

0.48

CopyWith

0.48

IsMutable

0.47

resave

0.47

 newfound

0.46

Activations Density 0.256%

references to previous relationships or past associations with the term "former."

No Comments

No Known Activations