INDEX

Explanations

references to nighttime events or activities

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

"night"

np_acts-logits-general · gemini-2.0-flash

night of or night's

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

 Públicas

-0.51

 alça

-0.42

dozer

-0.39

 Sprung

-0.37

 fãs

-0.37

 huom

-0.36

 femininos

-0.36

 cepillo

-0.36

balleur

-0.36

 faço

-0.35

POSITIVE LOGITS

Night

1.41

 night

1.41

 Night

1.38

 NIGHT

1.32

night

1.25

 Nights

1.14

 nights

1.13

Nights

1.13

NIGHT

1.08

 tonight

1.02

Activations Density 0.033%

references to nighttime events or activities

"night"

night of or night's

No Comments

No Known Activations