INDEX

Explanations

phrases conveying positive judgment or recommendation

oai_token-act-pair · gpt-3.5-turbo

expressions of positive evaluations or recommendations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alties

-0.71

anism

-0.71

naires

-0.66

Dialogue

-0.65

views

-0.65

AIDS

-0.64

uploads

-0.64

\<

-0.63

internal

-0.63

 Acad

-0.63

POSITIVE LOGITS

 great

1.26

 terrific

1.25

 fantastic

1.24

 worthwhile

1.24

 lifes

1.21

 nice

1.21

 wonderful

1.20

 valuable

1.13

 breeze

1.11

 boon

1.09

Activations Density 0.227%

phrases conveying positive judgment or recommendation

expressions of positive evaluations or recommendations

No Comments

No Known Activations