INDEX

Explanations

symbols indicating emphasis or importance

oai_token-act-pair · gpt-3.5-turbo

items or components related to cultural or social reviews

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

omorphic

-0.97

 cleaner

-0.83

conservancy

-0.78

hon

-0.74

 processed

-0.73

omial

-0.72

 arte

-0.71

 bonded

-0.69

 nesting

-0.68

emark

-0.68

POSITIVE LOGITS

âĵĺ

0.92

>[

0.84

 Ibid

0.82

quickShipAvailable

0.80

––

0.78

soDeliveryDate

0.77

Dialogue

0.76

while

0.75

paragraph

0.75

âĢ¢âĢ¢

0.74

Activations Density 0.022%

symbols indicating emphasis or importance

items or components related to cultural or social reviews

No Comments

No Known Activations