INDEX

Explanations

occurrences of the word "was"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

WebElementEntity

-0.53

esModule

-0.47

dropIfExists

-0.46

didSet

-0.46

Panamoan

-0.43

rungsseite

-0.43

makeConstraints

-0.43

Personendaten

-0.42

apimachinery

-0.41

Enllaces

-0.40

POSITIVE LOGITS

 told

0.59

 wondering

0.58

 pleasantly

0.57

 approached

0.52

 surprised

0.51

 disappointed

0.50

 amazed

0.50

 browsing

0.48

Told

0.47

 Told

0.47

Activations Density 0.265%

occurrences of the word "was"

No Comments

No Known Activations