INDEX

Explanations

numerical values and time measurements

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BeginContext

-0.79

styleType

-0.74

ListItemIcon

-0.69

UnusedPrivate

-0.68

 Савезне

-0.68

createCanvas

-0.68

subsubsection

-0.64

awtextra

-0.63

CloseOperation

-0.62

ResumeLayout

-0.62

POSITIVE LOGITS

;;)

0.50

 pinulongan

0.48

 météo

0.42

owulf

0.42

Etimo

0.41

 Utilizamos

0.41

$)

0.40

 jemanden

0.40

beleid

0.40

httphttps

0.40

Activations Density 1.737%

numerical values and time measurements

No Comments

No Known Activations