INDEX

Explanations

PG ratings

oai_token-act-pair · gpt-3.5-turbo

references to "PG" ratings or classifications in a context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ufact

-0.87

Ic

-0.76

joined

-0.71

puter

-0.71

 Bened

-0.64

fact

-0.64

 Bones

-0.63

cles

-0.62

bones

-0.62

igh

-0.62

POSITIVE LOGITS

raphic

0.94

allery

0.93

ENC

0.90

PI

0.90

ENE

0.90

asus

0.87

RAM

0.83

omez

0.81

reet

0.80

insula

0.80

Activations Density 0.014%

PG ratings

references to "PG" ratings or classifications in a context

No Comments

No Known Activations