INDEX

Explanations

markup tags and formatting indicators in text

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

mathematical notation with braces

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

 Taktlose

-0.59

AddTagHelper

-0.59

+#+#

-0.54

 MainAxisSize

-0.54

DIPSETTING

-0.48

ագրություններ

-0.47

SharedCtor

-0.45

 ویکی‌پدی

-0.45

 EClass

-0.45

UnitTesting

-0.44

POSITIVE LOGITS

hless

0.51

 martyr

0.50

Deck

0.45

paku

0.45

RTDA

0.44

 שוליים

0.44

ALK

0.44

betic

0.44

 defaultstate

0.44

 Deck

0.43

Activations Density 0.001%

markup tags and formatting indicators in text

mathematical notation with braces

No Comments

No Known Activations