INDEX

Explanations

hyphenated phrases that include the word "top"

oai_token-act-pair · gpt-3.5-turbo

terms related to rankings and levels of quality or importance

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

PsyNetMessage

-0.63

igham

-0.61

weak

-0.61

 Pengu

-0.61

coni

-0.60

Äĩ

-0.58

Love

-0.57

 Bonds

-0.56

ingham

-0.56

Clar

-0.55

POSITIVE LOGITS

yip

0.87

 earners

0.84

 contender

0.78

 swoop

0.65

tein

0.64

 notch

0.64

holder

0.64

seller

0.64

Thumbnail

0.63

 charts

0.63

Activations Density 0.159%

hyphenated phrases that include the word "top"

terms related to rankings and levels of quality or importance

No Comments

No Known Activations