INDEX

Explanations

actions or behaviors that are done willingly or voluntarily by individuals

oai_token-act-pair · gpt-3.5-turbo

terms related to voluntary actions or choices

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

 Said

-0.70

 Millennium

-0.69

ador

-0.68

 Horizons

-0.66

 Killer

-0.65

nar

-0.64

cas

-0.64

udeb

-0.63

AQ

-0.63

our

-0.63

POSITIVE LOGITS

 voluntarily

1.24

 willingly

1.09

 surrendered

1.03

 relinqu

1.01

 confess

0.92

 surrender

0.91

untarily

0.89

 sacrificing

0.86

 undergo

0.86

 compromising

0.86

Activations Density 0.006%

actions or behaviors that are done willingly or voluntarily by individuals

terms related to voluntary actions or choices

No Comments

No Known Activations