INDEX

Explanations

numbers and ratings in a list or structured format

oai_token-act-pair · gpt-3.5-turbo

numerical values associated with votes or scores

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Wiley

-0.77

iak

-0.69

Accessory

-0.68

assetsadobe

-0.65

 Rout

-0.65

 extrad

-0.64

agascar

-0.64

aunt

-0.63

¥µ

-0.63

imilar

-0.63

POSITIVE LOGITS

 votes

2.16

 vote

2.15

 Vote

2.06

 voting

2.06

vote

2.02

 Voting

1.95

 ballots

1.94

Vote

1.92

 Votes

1.92

votes

1.89

Activations Density 0.459%

numbers and ratings in a list or structured format

numerical values associated with votes or scores

No Comments

No Known Activations