INDEX

Explanations

words and phrases related to actions, particularly those that indicate a process or state of being

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

 anything

-1.74

 nothing

-1.67

 birth

-1.66

 shit

-1.55

 maternal

-1.52

 place

-1.46

 outgoing

-1.45

 functioning

-1.45

ught

-1.44

...?"

-1.44

POSITIVE LOGITS

naire

3.13

uclide

2.03

nier

2.00

etics

1.97

nel

1.95

berg

1.95

nement

1.84

eered

1.77

itrile

1.71

eria

1.68

Activations Density 0.163%

words and phrases related to actions, particularly those that indicate a process or state of being

No Comments

No Known Activations