INDEX

Explanations

words related to living spaces or residences, especially focusing on the word "dwell"

oai_token-act-pair · gpt-3.5-turbo

terms related to habitation or living arrangements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

xon

-0.79

 athlet

-0.72

ammy

-0.70

acha

-0.69

enic

-0.68

ripp

-0.66

apa

-0.64

ussian

-0.63

software

-0.62

alky

-0.62

POSITIVE LOGITS

 dwell

1.35

 Dwell

1.09

 dwelling

1.07

ers

0.93

erness

0.91

lords

0.91

erers

0.88

dw

0.86

rences

0.76

Discuss

0.72

Activations Density 0.008%

words related to living spaces or residences, especially focusing on the word "dwell"

terms related to habitation or living arrangements

No Comments

No Known Activations