INDEX

Explanations

numerical values appearing in a structured format

oai_token-act-pair · gpt-3.5-turbo

measurements and quantities in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

phrine

-0.94

 Halls

-0.82

theless

-0.73

deen

-0.66

 GOODMAN

-0.65

hower

-0.64

thirds

-0.63

 Galile

-0.62

 kitchens

-0.61

 Rumble

-0.60

POSITIVE LOGITS

icago

0.97

.,

0.95

./

0.82

 ........

0.81

ickr

0.80

Avg

0.78

hered

0.77

emp

0.77

nyder

0.76

iances

0.75

Activations Density 0.023%

numerical values appearing in a structured format

measurements and quantities in various contexts

No Comments

No Known Activations