INDEX

Explanations

dates in the format of year followed by a two-digit number

oai_token-act-pair · gpt-3.5-turbo

specific numerical dates and years

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

plom

-0.86

afort

-0.83

kell

-0.79

ablo

-0.75

Ã±o

-0.72

hammad

-0.70

pora

-0.70

hei

-0.70

hess

-0.69

etsk

-0.69

POSITIVE LOGITS

âĸĪâĸĪ

0.92

committee

0.82

SW

0.81

th

0.79

%"

0.75

mph

0.74

ishers

0.73

ISH

0.72

0.71

Activations Density 0.045%

dates in the format of year followed by a two-digit number

specific numerical dates and years

No Comments

No Known Activations