INDEX

Explanations

the prefix "non-" used to form negations

oai_token-act-pair · gpt-4-turbo

words related to non-official or non-profit entities

oai_token-act-pair · gpt-3.5-turbo

references to non-official entities or accounts in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ulhu

-0.96

ħĭ

-0.82

 Dickinson

-0.79

HCR

-0.78

ĸļ

-0.77

DragonMagazine

-0.73

=#

-0.71

iage

-0.71

ometimes

-0.70

 Twain

-0.69

POSITIVE LOGITS

level

1.07

sized

1.06

legged

1.04

exclusive

1.03

issue

1.02

conscious

0.96

specific

0.95

related

0.95

based

0.94

function

0.94

Activations Density 0.378%

the prefix "non-" used to form negations

words related to non-official or non-profit entities

references to non-official entities or accounts in various contexts

No Comments

No Known Activations