INDEX

Explanations

commands or programming instructions

oai_token-act-pair · gpt-3.5-turbo

technical terms and jargon related to organization and formatting

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

 Bard

-0.74

ILS

-0.74

VAT

-0.72

 Skywalker

-0.72

 Olive

-0.70

OTS

-0.70

Lens

-0.68

Tik

-0.68

 Butterfly

-0.66

Elk

-0.66

POSITIVE LOGITS

plement

0.95

nex

0.93

thur

0.92

rha

0.91

mediate

0.89

posing

0.87

rupt

0.84

pill

0.84

puter

0.83

nant

0.83

Activations Density 0.033%

commands or programming instructions

technical terms and jargon related to organization and formatting

No Comments

No Known Activations