INDEX

Explanations

text related to documentation or recording information

oai_token-act-pair · gpt-3.5-turbo

mentions of "document" and its derivatives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

icidal

-0.70

oppers

-0.68

bane

-0.67

Ĩ

-0.67

 swings

-0.65

 slopes

-0.63

Stars

-0.63

ÃŁ

-0.62

ptions

-0.61

uno

-0.61

POSITIVE LOGITS

 document

3.73

document

2.37

 documents

2.36

 Document

2.10

Document

2.03

 Documents

1.82

 documentation

1.60

 memo

1.56

Documents

1.56

 memorandum

1.41

Activations Density 0.014%

text related to documentation or recording information

mentions of "document" and its derivatives

No Comments

No Known Activations