INDEX

Explanations

references to video content and media

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

"video"

np_acts-logits-general · gemini-2.0-flash

video

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

 mijne

-0.49

 zelve

-0.49

 çat

-0.48

 verster

-0.47

 trouw

-0.47

 ouder

-0.45

 trưng

-0.44

มาะ

-0.44

 lijn

-0.44

 geluk

-0.43

POSITIVE LOGITS

 video

1.18

 Video

1.16

Video

1.16

video

1.11

 videos

1.06

 Videos

1.05

 VIDEO

1.03

Videos

0.93

videos

0.93

VIDEO

0.91

Activations Density 0.055%

references to video content and media

"video"

video

No Comments

No Known Activations