INDEX

Explanations

vote-related information such as vote counts and identifiers

oai_token-act-pair · gpt-3.5-turbo

elements related to rankings or competitions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

etheless

-0.65

inarily

-0.65

Ĥª

-0.60

notations

-0.54

 GOODMAN

-0.53

enei

-0.53

 Canaver

-0.52

 atmosp

-0.52

 tragically

-0.51

idelines

-0.50

POSITIVE LOGITS

->

0.70

[-

0.67

(-

0.65

(+

0.64

0.60

 Allows

0.59

Allows

0.58

=>

0.58

["

0.58

Activations Density 0.592%

vote-related information such as vote counts and identifiers

elements related to rankings or competitions

No Comments

No Known Activations