INDEX

Explanations

legal or court-related phrases, particularly those about judicial opinions or citations

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

ibase

-1.78

ivos

-1.73

pective

-1.71

ÃŃas

-1.70

aurus

-1.68

STA

-1.64

ARGS

-1.63

ista

-1.59

enses

-1.59

ense

-1.58

POSITIVE LOGITS

 suspicion

1.75

woke

1.74

↵

1.64

↵

1.64

<|outofrange|>

1.64

↵            ↵

1.64

<|outofrange|>

1.64

č↵

1.64

↵

1.64

Activations Density 0.371%

legal or court-related phrases, particularly those about judicial opinions or citations

No Comments

No Known Activations