INDEX

Explanations

information related to announcements and releases

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

releases and announcements

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

WriteTagHelper

-0.51

AddTagHelper

-0.45

 tartalomajánló

-0.45

 informée

-0.42

})}

-0.40

">{{$

-0.39

damento

-0.38

 samochód

-0.36

Diweddarwch

-0.36

 psicología

-0.36

POSITIVE LOGITS

 initComponents

0.54

 teased

0.52

 releases

0.51

revealed

0.50

 teaser

0.50

 bloggers

0.49

 teasing

0.48

released

0.48

 Grimal

0.48

substituted

0.47

Activations Density 0.007%

information related to announcements and releases

releases and announcements

No Comments

No Known Activations