INDEX

Explanations

terms related to social contexts and relationships

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/1-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.1.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

")]

-1.88

 induced

-1.63

 stratified

-1.54

 supplemented

-1.52

uchs

-1.51

sembles

-1.50

*](#

-1.49

 detected

-1.46

uded

-1.42

""

-1.41

POSITIVE LOGITS

burgh

1.86

voice

1.84

hurst

1.65

hearted

1.59

bay

1.57

love

1.53

stown

1.50

áĢº

1.44

wick

1.40

 Sisters

1.39

Activations Density 0.892%

terms related to social contexts and relationships

No Comments

No Known Activations