INDEX

Explanations

phrases expressing specific legal or procedural contexts

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-,

-0.33

 fodder

-0.32

RenderAtEndOf

-0.32

Siehe

-0.31

ürd

-0.31

emer

-0.31

ticular

-0.30

 vaisselle

-0.30

Palmarès

-0.30

 futures

-0.30

POSITIVE LOGITS

awtextra

0.68

 autorytatywna

0.66

цездатний

0.64



0.64

SPJ

0.63

rungsseite

0.62

SourceChecksum

0.61

 мәкал

0.61

✨:

0.59

 Италијани

0.59

Activations Density 0.102%

phrases expressing specific legal or procedural contexts

No Comments

No Known Activations