INDEX

Explanations

indications of slight amounts or nuances

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a moderate quantity or degree

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

ħĭ

-0.90

rat

-0.79

ORTS

-0.76

rats

-0.74

oons

-0.70

brates

-0.68

acus

-0.66

apons

-0.66

ARS

-0.65

ESE

-0.64

POSITIVE LOGITS

 luck

0.88

fun

0.76

icial

0.76

 extra

0.71

 consolation

0.68

 juicy

0.68

 sidx

0.65

 juice

0.65

ainment

0.65

 brilliance

0.63

Activations Density 0.064%

indications of slight amounts or nuances

phrases that indicate a moderate quantity or degree

No Comments

No Known Activations