INDEX

Explanations

terms related to future events or predictions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 poffible

-0.56

 laſſen

-0.55

vician

-0.55

 Infórmanos

-0.52

 himſelf

-0.51

 betweenstory

-0.49

phazard

-0.49

↹

-0.49

 indígen

-0.49

 increí

-0.48

POSITIVE LOGITS

 مشين

0.56

ViewImports

0.42

 Мексичка

0.34

 Biôgrafia

0.33

Referenties

0.32

BASEPATH

0.31

Personendaten

0.30

 Seasons

0.30

 MaterialApp

0.29

PerSecond

0.29

Activations Density 0.108%

terms related to future events or predictions

No Comments

No Known Activations