INDEX

Explanations

structural forms or formats typically associated with documentation or content

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĻĤ

-2.67

·

-2.37

Ļª

-2.15

±

-2.15

§

-2.12

ĸ´

-2.03

¯

-2.03

Ń

-1.96

·¸

-1.94

ľĵ

-1.94

POSITIVE LOGITS

idable

2.57

ality

1.99

ally

1.97

ably

1.80

ics

1.80

nel

1.79

ontal

1.79

ulating

1.76

Tables

1.72

ibus

1.71

Activations Density 0.017%

structural forms or formats typically associated with documentation or content

No Comments

No Known Activations