INDEX

Explanations

phrases that emphasize the concepts of allowing, letting, and flow within various contexts

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

allowing actions to continue

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

essoal

-0.42

 توا

-0.40

 jäl

-0.39

 Vort

-0.39

 juuri

-0.39

TagHelper

-0.38

OPTS

-0.38

estimm

-0.38

 opdracht

-0.37

DELETED

-0.35

POSITIVE LOGITS

 Laissez

0.62

 poveznice

0.56

 laissez

0.56

 ligiloj

0.55

存于互联网档案馆

0.54

let

0.54

 linkovi

0.51

 CURIAM

0.50

setViewportView

0.49

 allow

0.48

Activations Density 0.045%

phrases that emphasize the concepts of allowing, letting, and flow within various contexts

allowing actions to continue

No Comments

No Known Activations