INDEX

Explanations

mentions of the first occurrence or achievement of something in a series or career

oai_token-act-pair · gpt-3.5-turbo

references to first-time events or significant appearances

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

riber

-0.75

ensional

-0.72

 abound

-0.72

isson

-0.68

orius

-0.66

 Berman

-0.66

 Registrar

-0.64

 Archives

-0.63

rupulous

-0.62

bly

-0.62

POSITIVE LOGITS

 outing

1.01

 foray

0.89

 EVER

0.83

 outings

0.82

 ever

0.78

Ever

0.78

 imaginable

0.71

 counterparts

0.71

mates

0.70

 appearance

0.68

Activations Density 0.267%

mentions of the first occurrence or achievement of something in a series or career

references to first-time events or significant appearances

No Comments

No Known Activations