INDEX

Explanations

specific characters or formatting elements in a document

oai_token-act-pair · gpt-3.5-turbo

patterns involved in lists or groupings of items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hement

-0.74

 destro

-0.66

 citiz

-0.66

 speeding

-0.66

 morphed

-0.63

 Sach

-0.61

 eroded

-0.61

 Thornton

-0.60

 Singh

-0.59

 ende

-0.59

POSITIVE LOGITS

BuyableInstoreAndOnline

0.79

Products

0.79

Unknown

0.74

Height

0.73

Interstitial

0.73

Spoiler

0.72

SCP

0.71

Ey

0.71

Ingredients

0.71

rossover

0.70

Activations Density 0.106%

specific characters or formatting elements in a document

patterns involved in lists or groupings of items

No Comments

No Known Activations