INDEX

Explanations

conversational phrases and prompts indicating a shift in topic or direction in the text

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

now turning to a topic

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Paglinawan

-0.58

 setUsername

-0.56

tonode

-0.56

 ainfi

-0.54

 setCode

-0.48

səhifə

-0.47

 queſta

-0.47

 intStringLen

-0.46

peated

-0.46

eseorang

-0.46

POSITIVE LOGITS

VersionUID

0.50

ContentAsync

0.40

forName

0.39

consulté

0.39

 crux

0.38

 problem

0.36

说说

0.34

問題は

0.34

конец

0.34

 matters

0.34

Activations Density 0.040%

conversational phrases and prompts indicating a shift in topic or direction in the text

now turning to a topic

No Comments

No Known Activations