INDEX

Explanations

proper nouns with a specific pattern

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names

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

 predec

-0.83

 cryst

-0.69

avorite

-0.68

pection

-0.65

 Thumbnails

-0.64

ishable

-0.64

sufficient

-0.63

theless

-0.61

 artif

-0.61

 bulky

-0.61

POSITIVE LOGITS

 Rowling

0.85

EStream

0.74

zyk

0.70

schild

0.69

pod

0.69

nik

0.69

 Sabha

0.68

ernaut

0.68

xon

0.67

bush

0.66

Activations Density 0.397%

proper nouns with a specific pattern

proper nouns and names

No Comments

No Known Activations