INDEX

Explanations

references to success or positive outcomes in a programming or user interface context

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

setVerticalGroup

-0.68

onedDateTime

-0.55

 createState

-0.52

IntoConstraints

-0.50

rungsseite

-0.50

BagConstraints

-0.50

ⓧ

-0.47

QMetaType

-0.46

endpush

-0.46

 utafitiHapana

-0.43

POSITIVE LOGITS

각주

0.47

 victor

0.45

 warn

0.44

 warning

0.44

ParallelGroup

0.42

âtre

0.41

 alert

0.40

TTS

0.40

THOOK

0.40

alerts

0.39

Activations Density 0.000%

references to success or positive outcomes in a programming or user interface context

No Comments

No Known Activations