INDEX

Explanations

legal terminology and references to court rulings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

+:+

-0.69

InSection

-0.68

цездатний

-0.61

verwijspagina

-0.57

Personensuche

-0.57

 ویکی‌پدیای

-0.57

 Normdatei

-0.55

DispatchToProps

-0.53

forRoot

-0.52

 QWidget

-0.52

POSITIVE LOGITS

ornment

0.54

hability

0.53

 Perce

0.51

versaries

0.51

Населення

0.50

 sympy

0.49

AndEndTag

0.49

isière

0.48

biam

0.47

Shh

0.47

Activations Density 0.109%

legal terminology and references to court rulings

No Comments

No Known Activations