INDEX

Explanations

occurrences of labels in structured formats, such as programming code or data representations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

extrême

-0.52

 Milne

-0.48

Spoljašnje

-0.47

üyor

-0.47

orth

-0.46

 mord

-0.46

 Linn

-0.46

 Burgos

-0.45

Xp

-0.45

niksi

-0.44

POSITIVE LOGITS

getLabel

0.63

tiel

0.58

ophageal

0.57

 ricevere

0.56

orteur

0.55

LABEL

0.55

jenigen

0.54

 Label

0.54

thel

0.53

SharedCtor

0.53

Activations Density 0.493%

occurrences of labels in structured formats, such as programming code or data representations

No Comments

No Known Activations