INDEX

Explanations

sections containing structured data or lists

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.0.hook_mlp_out

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

 same

-1.58

 prong

-1.52

 cardinal

-1.48

 aloud

-1.48

 polyclonal

-1.44

 foreground

-1.43

 intimately

-1.42

 allele

-1.41

 vigorously

-1.40

 temper

-1.40

POSITIVE LOGITS

ĻĤ

2.67

ħ

2.50

»¿

2.48

»

2.38

ĺ

2.28

¥

2.24

ĥ½

2.18

¨

2.15

Ī

2.08

Īĺ

2.06

Activations Density 1.604%

sections containing structured data or lists

No Comments

No Known Activations