INDEX

Explanations

words related to physical sensations or actions

oai_token-act-pair · gpt-3.5-turbo

words associated with names or identities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 triangle

-0.73

 absence

-0.71

ICAN

-0.68

 inaction

-0.67

GBT

-0.66

 wealth

-0.65

ĪĴ

-0.64

church

-0.63

 staging

-0.62

UCT

-0.62

POSITIVE LOGITS

glers

0.90

oslav

0.80

vous

0.74

dos

0.73

onyms

0.71

puff

0.70

azines

0.70

 Picks

0.70

bilt

0.70

tty

0.70

Activations Density 0.048%

words related to physical sensations or actions

words associated with names or identities

No Comments

No Known Activations