INDEX

Explanations

commands or directives aimed at others

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

telling instructions to someone

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 Neub

-0.41

 halten

-0.40

כז

-0.39

rijf

-0.36

็ง

-0.36

கு

-0.36

cpy

-0.35

 Marsh

-0.35

Marcel

-0.34

colazione

-0.34

POSITIVE LOGITS

Datuak

0.64

 utafitiHapana

0.59

 propOrder

0.56

 EconPapers

0.55

Geplaatst

0.54

󠁢

0.54

Hochspringen

0.53

oa̍t

0.50

 ligiloj

0.48

aarrggbb

0.48

Activations Density 0.048%

commands or directives aimed at others

telling instructions to someone

No Comments

No Known Activations