INDEX

Explanations

references to categories or classifications within texts

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

long s words

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

 AssemblyCompany

-0.59

LookAnd

-0.57

 AssemblyTitle

-0.56

تقاوى

-0.54

 AssemblyVersion

-0.52

RenderAtEndOf

-0.51

dafx

-0.50

 Autorisations

-0.50

 useAppContext

-0.48

 Signalez

-0.46

POSITIVE LOGITS

 itſelf

0.47

wiſe

0.46

 ſel

0.46

bucha

0.46

ctio

0.45

sohn

0.44

ſelf

0.44

Mort

0.43

Salad

0.43

ophagus

0.42

Activations Density 0.042%

references to categories or classifications within texts

long s words

No Comments

No Known Activations