INDEX

Explanations

numbers related to quantities or counts

oai_token-act-pair · gpt-3.5-turbo

numerical expressions indicating quantities or counts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

uca

-0.68

ESE

-0.66

tick

-0.64

ask

-0.64

isu

-0.64

artz

-0.63

illard

-0.62

ettel

-0.60

answer

-0.59

agame

-0.59

POSITIVE LOGITS

 others

1.75

 other

1.56

other

1.25

 Others

1.22

 else

1.19

 accompanying

1.15

Others

1.12

 adjoining

1.09

OTHER

1.09

 OTHER

1.06

Activations Density 0.270%

numbers related to quantities or counts

numerical expressions indicating quantities or counts

No Comments

No Known Activations