INDEX

Explanations

References to time lapses or microphotography

oai_token-act-pair · gpt-3.5-turbo

terms related to temporal changes or sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

cu

-0.75

 pillow

-0.70

 suitcase

-0.68

 purse

-0.66

Lov

-0.65

 parks

-0.64

 ransom

-0.63

 bount

-0.63

 cursed

-0.62

 tossed

-0.62

POSITIVE LOGITS

apse

1.54

apses

1.35

arella

0.99

avorite

0.94

edIn

0.94

Clock

0.91

ideos

0.90

oldemort

0.85

asar

0.85

hetamine

0.84

Activations Density 0.011%

References to time lapses or microphotography

terms related to temporal changes or sequences

No Comments

No Known Activations