INDEX

Explanations

sequences of numbers or data patterns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

 Füßen

-0.53

 ujednoznacz

-0.49

 desmotivaciones

-0.47

 herida

-0.47

 wußte

-0.47

 cortos

-0.46

 Jerusalén

-0.46

 Grüsse

-0.46

 Absicht

-0.45

 sostu

-0.45

POSITIVE LOGITS

Personendaten

0.56

CloseOperation

0.52

mybatisplus

0.46

SizeMode

0.43

 Spinning

0.42

phorbia

0.42

sembles

0.42

 AppCompatTheme

0.42

DOUT

0.41

DeleteBehavior

0.41

Activations Density 0.553%

sequences of numbers or data patterns

No Comments

No Known Activations