INDEX

Explanations

dates expressed as the day of the week

oai_token-act-pair · gpt-3.5-turbo

the mention of the day "Friday."

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

ership

-1.00

hip

-0.82

vironment

-0.77

ports

-0.77

atche

-0.77

ris

-0.76

respons

-0.75

abet

-0.74

colour

-0.73

lessly

-0.73

POSITIVE LOGITS

 morning

1.41

 afternoon

1.39

 night

1.25

 evening

1.19

 mornings

1.16

 Night

1.16

 nights

1.04

Night

0.92

 evenings

0.89

 dusk

0.84

Activations Density 0.026%

dates expressed as the day of the week

the mention of the day "Friday."

No Comments

No Known Activations