INDEX

Explanations

indications of threats and attacks, particularly in a context involving groups or individuals that may cause harm

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

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

цездатний

-0.47

hoeddwyd

-0.46

❋

-0.46

Personensuche

-0.44

SaveChangesAsync

-0.42

 AssemblyCompany

-0.41

//

-0.41

ngdoc

-0.39

disposing

-0.39

azul

-0.38

POSITIVE LOGITS

abases

0.54

 invade

0.53

 invading

0.46

 attack

0.45

 invasion

0.45

 exploiting

0.44

 threatening

0.43

 assaulting

0.43

uric

0.43

abus

0.43

Activations Density 0.068%

indications of threats and attacks, particularly in a context involving groups or individuals that may cause harm

No Comments

No Known Activations