INDEX

Explanations

phrases expressing confusion or seeking clarification

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

ThroughAttribute

-0.65

Datuak

-0.56

Vidite

-0.54

ftagPool

-0.50

 Paglinawan

-0.50

دانشنامهٔ

-0.49

 autorytatywna

-0.48



-0.46

StructEnd

-0.46

下载附件

-0.46

POSITIVE LOGITS

意思

0.40

queta

0.40

 cera

0.39

Example

0.38

ocumented

0.38

 anormal

0.38

CreateModel

0.37

 vaguely

0.37

NullException

0.37

Anyway

0.37

Activations Density 0.051%

phrases expressing confusion or seeking clarification

No Comments

No Known Activations