INDEX

Explanations

sequences or patterns of numbers and symbols, often related to codes or identifiers

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

legal markers like AKA or Appellee

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

windowFixed

-0.46

polo

-0.43

cnia

-0.42

 ExecuteAsync

-0.41

ergies

-0.40

 فريبيس

-0.40

bog

-0.40

-0.39

 umge

-0.39

 Mallory

-0.39

POSITIVE LOGITS

!*\

0.52

AndEndTag

0.50

AccessorTable

0.45

ChrTalk

0.42

 htons

0.41

ьаж

0.40

KURZBESCHREIBUNG

0.39

:✨

0.38

granate

0.36

 règne

0.36

Activations Density 0.059%

sequences or patterns of numbers and symbols, often related to codes or identifiers

legal markers like AKA or Appellee

No Comments

No Known Activations