INDEX

Explanations

punctuation marks and phrases that indicate reading or buying recommendations

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

previous or existing

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

+#+#

-0.63

 Signalez

-0.63

⤹

-0.59

AddHtmlAttribute

-0.55

AISSEE

-0.55

bootstrapcdn

-0.54

CppMethod

-0.54

الحياه

-0.54

 BoxFit

-0.53

tanleria

-0.53

POSITIVE LOGITS

 previous

0.40

openzeppelin

0.38

 precedent

0.36

comp

0.33

 Wikime

0.33

 existing

0.33

 precedente

0.33

 asim

0.33

 precedenti

0.31

 ProtoMessage

0.31

Activations Density 0.017%

punctuation marks and phrases that indicate reading or buying recommendations

previous or existing

No Comments

No Known Activations