INDEX

Explanations

the particle "を" (wo), which indicates the object of a sentence in Japanese

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_1/width_16k/average_l0_102

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 purpoſe

-1.02

 houſe

-0.95

 Jefus

-0.94

 itſelf

-0.92

 pleaſure

-0.92

 Chriftian

-0.91

 myſelf

-0.89

 Chriſt

-0.89

 Monfieur

-0.88

 iſt

-0.87

POSITIVE LOGITS

を

2.55

를

2.00

을

1.93

を

1.59

를

1.38

いを

1.36

을

1.35

子を

1.32

りを

1.30

曲を

1.20

Activations Density 0.018%

the particle "を" (wo), which indicates the object of a sentence in Japanese

No Comments

No Known Activations