INDEX

Explanations

code snippets with ellipses indicating omitted code

oai_token-act-pair · gpt-3.5-turbo

ellipses or indications of omitted text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

entimes

-0.85

 Lomb

-0.74

onite

-0.71

grass

-0.70

 animosity

-0.68

 contrace

-0.67

 confidentiality

-0.67

 nonviolent

-0.67

 pockets

-0.66

 suspic

-0.65

POSITIVE LOGITS

ãĥ¯ãĥ³

0.76

/>

0.73

BUT

0.71

ãĥĺ

0.69

soDeliveryDate

0.68

 âĢº

0.68

 Retrieved

0.67

Done

0.67

SK

0.63

CSS

0.63

Activations Density 0.045%

code snippets with ellipses indicating omitted code

ellipses or indications of omitted text

No Comments

No Known Activations