INDEX

Explanations

expressions related to capabilities and actions

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

 bocca

-0.43

 calcetines

-0.42

charCodeAt

-0.41

ACIONAL

-0.40

MessageState

-0.40

yarnpkg

-0.39

+:+

-0.37

 pungkas

-0.37

limiento

-0.37

esistenza

-0.36

POSITIVE LOGITS

 pinulongan

0.57

 ویکی‌پدیا

0.53

dible

0.50

 Accessible

0.47

 doable

0.47

 reproducible

0.46

 transfieras

0.46

 digestible

0.45

OGND

0.45

createCell

0.45

Activations Density 0.064%

expressions related to capabilities and actions

No Comments

No Known Activations