INDEX

Explanations

numerical patterns or codes in text data

oai_token-act-pair · gpt-3.5-turbo

structured data labels or identifiers

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

iasm

-0.82

oldown

-0.82

acity

-0.78

jri

-0.73

ocr

-0.72

 toppled

-0.72

gross

-0.71

 tears

-0.71

uracy

-0.70

phthal

-0.70

POSITIVE LOGITS

VIS

0.81

Anonymous

0.77

 Notification

0.73

 SOLD

0.72

 Generic

0.70

 Quantity

0.70

 Unknown

0.69

 INFORMATION

0.69

 CARD

0.68

 Cosponsors

0.68

Activations Density 0.095%

numerical patterns or codes in text data

structured data labels or identifiers

No Comments

No Known Activations