INDEX

Explanations

phrases conveying positive sentiment or approval

oai_token-act-pair · gpt-3.5-turbo

positive affirmations or commendatory phrases

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

Downloadha

-0.74

abel

-0.74

minent

-0.72

ËĪ

-0.69

IDES

-0.69

 Strauss

-0.69

agin

-0.68

actionDate

-0.68

resent

-0.67

URA

-0.65

POSITIVE LOGITS

job

1.18

 luck

1.12

bye

1.08

 timing

1.06

 thing

1.03

 grief

0.96

 idea

0.94

 heavens

0.94

 gracious

0.92

bye

0.91

Activations Density 0.093%

phrases conveying positive sentiment or approval

positive affirmations or commendatory phrases

No Comments

No Known Activations