INDEX

Explanations

words related to alerts or cautions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "warning."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

meric

-0.71

 Abram

-0.70

opus

-0.67

cles

-0.66

competitive

-0.66

akeru

-0.65

thereal

-0.64

berman

-0.64

aepernick

-0.63

ovy

-0.62

POSITIVE LOGITS

 warning

3.98

 warnings

2.89

warning

2.77

 Warning

2.59

 warn

2.15

Warning

1.99

 WARNING

1.92

warn

1.88

 Warn

1.85

 caution

1.85

Activations Density 0.030%

words related to alerts or cautions

instances of the word "warning."

No Comments

No Known Activations