INDEX

Explanations

elements related to investigations and evidence in legal or medical contexts

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

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_32k/average_l0_76

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

float32

Hook Name

blocks.12.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

GraphicsUnit

-0.86

awtextra

-0.71

 linkovi

-0.68

InstrumentedTest

-0.65

 BorderRadius

-0.64

 incomplète

-0.62

 ligiloj

-0.61

 resourceCulture

-0.61

oznam

-0.61

 ويكيميديا

-0.60

POSITIVE LOGITS

 anyway

0.59

anyway

0.57

 regardless

0.56

 Regardless

0.54

Regardless

0.53

 anyways

0.51

 Anyway

0.50

Anyway

0.50

UBE

0.50

 Toujours

0.49

Activations Density 0.295%

elements related to investigations and evidence in legal or medical contexts

No Comments

No Known Activations