INDEX

Explanations

self-described or self-proclaimed individuals or groups

oai_token-act-pair · gpt-3.5-turbo

phrases that categorize groups based on their self-identification

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ITNESS

-0.71

 ILCS

-0.70

INTER

-0.66

EMS

-0.65

EM

-0.65

Cu

-0.65

TPS

-0.64

SAF

-0.64

 Surv

-0.63

 Oaks

-0.63

POSITIVE LOGITS

proclaimed

0.98

described

0.92

tarian

0.92

identified

0.91

gettable

0.90

TextColor

0.90

ItemTracker

0.88

sty

0.87

etheless

0.86

quartered

0.84

Activations Density 0.025%

self-described or self-proclaimed individuals or groups

phrases that categorize groups based on their self-identification

No Comments

No Known Activations