INDEX

Explanations

references to first appearances or introductions of something, often in the context of debuts

oai_token-act-pair · gpt-3.5-turbo

references to debuts across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

enough

-0.65

learn

-0.63

asus

-0.63

hate

-0.61

fax

-0.61

Downloadha

-0.60

akia

-0.60

 Canaver

-0.58

thia

-0.57

pe

-0.55

POSITIVE LOGITS

antes

1.22

ante

1.15

ant

0.95

ants

0.92

antly

0.88

 episode

0.80

ary

0.74

edIn

0.72

iator

0.72

tained

0.70

Activations Density 0.041%

references to first appearances or introductions of something, often in the context of debuts

references to debuts across various contexts

No Comments

No Known Activations