INDEX

Explanations

phrases indicating actions or events related to needing, doing, or having experiences

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

"do" followed by a verb

np_acts-logits-general · gemini-2.0-flash

do / did + verb

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

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

Fp

-0.39

TRO

-0.37

dem

-0.36

ViewImports

-0.35

 vola

-0.35

 Thom

-0.34

Monday

-0.33

 Crawford

-0.33

כי

-0.33

 cantor

-0.33

POSITIVE LOGITS

erdings

0.71

 niestety

0.71

 indeed

0.71

ではある

0.70

 memang

0.68

anmoins

0.68

равда

0.68

 faktiskt

0.68

 fakty

0.67

 tatsächlich

0.65

Activations Density 0.200%

phrases indicating actions or events related to needing, doing, or having experiences

"do" followed by a verb

do / did + verb

No Comments

No Known Activations