INDEX

Explanations

mentions of specific actions involving physical force or harm

oai_token-act-pair · gpt-3.5-turbo

punctuation and phrases that indicate dialogue or testimony

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

atos

-0.68

license

-0.66

ciples

-0.66

hazard

-0.64

arius

-0.63

doors

-0.63

tesy

-0.61

ramid

-0.61

ertodd

-0.61

atus

-0.60

POSITIVE LOGITS

she

1.76

Ms

1.74

she

1.73

 hers

1.62

her

1.54

She

1.51

She

1.47

SHE

1.43

Mrs

1.40

Ms

1.36

Activations Density 0.711%

mentions of specific actions involving physical force or harm

punctuation and phrases that indicate dialogue or testimony

No Comments

No Known Activations