INDEX

Explanations

instances of the word "flex" and its variations, indicating a focus on flexibility or adaptability

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

sequences starting specific context

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

 EventArgs

-0.25

vid

-0.23

付

-0.23

kav

-0.21

||

-0.21

Ill

-0.21

 رو

-0.21

fortawesome

-0.21

õ

-0.20

POSITIVE LOGITS

rungsseite

0.81

intios

0.73

 MainAxisSize

0.68

 EconPapers

0.66

RTSC

0.64

oredCriteria

0.63

PostInfinity

0.63

LabelTagHelper

0.62

Personensuche

0.62

rela

0.59

Activations Density 0.000%

instances of the word "flex" and its variations, indicating a focus on flexibility or adaptability

sequences starting specific context

No Comments

No Known Activations