INDEX

Explanations

locations or positions mentioned relative to a point of reference

oai_token-act-pair · gpt-3.5-turbo

references to items located beneath other content

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

ledged

-0.72

ebook

-0.61

merce

-0.58

ache

-0.57

stadt

-0.56

sense

-0.55

ointment

-0.55

hler

-0.55

arl

-0.54

ident

-0.54

POSITIVE LOGITS

 below

3.53

below

2.70

 above

2.48

 Below

2.36

 BELOW

2.19

above

1.99

Below

1.93

 Above

1.66

 beneath

1.63

Above

1.42

Activations Density 0.017%

locations or positions mentioned relative to a point of reference

references to items located beneath other content

No Comments

No Known Activations