INDEX

Explanations

phrases related to specific individuals, potentially involving professional or personal information about them

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ran."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ready

-0.66

lda

-0.60

GMT

-0.60

OHN

-0.59

earable

-0.59

cember

-0.59

 Foley

-0.58

ENS

-0.58

 minors

-0.57

ãĤī

-0.56

POSITIVE LOGITS

vier

1.11

igans

0.96

igan

0.88

tenance

0.87

ormal

0.83

fo

0.83

thood

0.83

iety

0.83

quez

0.83

thus

0.82

Activations Density 0.034%

phrases related to specific individuals, potentially involving professional or personal information about them

instances of the word "ran."

No Comments

No Known Activations