INDEX

Explanations

references to products and branding within content

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vereign

-1.62

 infringement

-1.58

icates

-1.58

otide

-1.54

inary

-1.53

rapeutics

-1.51

portation

-1.49

olation

-1.46

osp

-1.44

inates

-1.42

POSITIVE LOGITS

±

2.29

ı

2.15

³

2.12

¦

2.08

°

2.00

»¿

2.00

Ŀ

1.98

ª

1.96

ĻĤ

1.95

ľ

1.95

Activations Density 3.403%

references to products and branding within content

No Comments

No Known Activations