INDEX

Explanations

the word "used" followed by either a numerical value or a verb

oai_token-act-pair · gpt-3.5-turbo

phrases indicating habitual actions or acclimatization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

oret

-0.63

 clinic

-0.62

 Buckingham

-0.61

 unfinished

-0.61

 mean

-0.59

 deviations

-0.58

 Baron

-0.57

 Cunningham

-0.57

 provisional

-0.57

Bin

-0.56

POSITIVE LOGITS

oise

0.83

 aback

0.80

illusion

0.79

ivated

0.78

apolis

0.77

inking

0.76

ient

0.75

ivating

0.75

atically

0.74

ience

0.74

Activations Density 0.165%

the word "used" followed by either a numerical value or a verb

phrases indicating habitual actions or acclimatization

No Comments

No Known Activations