INDEX

Explanations

locations mentioned in the text related to New Jersey

oai_token-act-pair · gpt-3.5-turbo

references to places, particularly related to Newark and New Jersey

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ware

-0.74

acies

-0.72

quer

-0.66

 Hels

-0.65

atable

-0.64

Ðº

-0.63

nered

-0.63

yip

-0.63

 ener

-0.63

arest

-0.63

POSITIVE LOGITS

 Devils

1.00

NJ

0.95

 Rutgers

0.92

 Yards

0.89

 Transit

0.86

shire

0.79

 Advance

0.72

 Newark

0.72

ocamp

0.71

 trustee

0.70

Activations Density 0.029%

locations mentioned in the text related to New Jersey

references to places, particularly related to Newark and New Jersey

No Comments

No Known Activations