INDEX

Explanations

numerical codes typically associated with publications or communications

oai_token-act-pair · gpt-3.5-turbo

numerical values and references to ISBNs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hare

-0.66

conserv

-0.62

ucha

-0.61

upon

-0.60

 signs

-0.59

andestine

-0.58

brace

-0.56

lest

-0.56

 flares

-0.56

 scissors

-0.56

POSITIVE LOGITS

 CrossRef

0.70

CLR

0.63

 istg

0.63

çīĪ

0.62

 Elon

0.61

Syl

0.61

enos

0.60

GD

0.60

 Huang

0.59

 attRot

0.59

Activations Density 0.133%

numerical codes typically associated with publications or communications

numerical values and references to ISBNs

No Comments

No Known Activations