INDEX

Explanations

segments related to educational content and learning activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 Italijanski

-0.57

 linkovi

-0.55

 дописавши

-0.54

Cyfeiriadau

-0.53

 صوتيه

-0.53

verwijspagina

-0.52

writeFieldEnd

-0.52

 resourceCulture

-0.52

 ब्रेकडाउन

-0.52

 Reisedaten

-0.50

POSITIVE LOGITS

 learning

0.56

 lesson

0.55

 teaching

0.52

 Learning

0.49

lesson

0.46

Lesson

0.45

 Lesson

0.44

 appren

0.43

Learning

0.43

 Learn

0.42

Activations Density 0.108%

segments related to educational content and learning activities

No Comments

No Known Activations