INDEX

Explanations

occurrences of the word "in"

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

python code execution or import

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

ſelves

-0.60

makeText

-0.52

Попис

-0.51

ⓧ

-0.48

unnitel

-0.47

 AspNetCore

-0.46

findpost

-0.45

 ſtill

-0.45

 whoſe

-0.45

 themſelves

-0.45

POSITIVE LOGITS

ValueStyle

0.68

PyExc

0.55

↴

0.54

toThrow

0.50

 مشين

0.50

 tartalomajánló

0.47

newArrayList

0.47

RTCF

0.46

extAlignment

0.46

)_/¯

0.46

Activations Density 0.050%

occurrences of the word "in"

python code execution or import

No Comments

No Known Activations