INDEX

Explanations

questions and commands

oai_token-act-pair · gpt-3.5-turbo

questions and inquiries starting with "What" or "Why."

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

Iv

-0.78

tin

-0.76

fm

-0.73

-0.69

boat

-0.68

tur

-0.68

åĲ

-0.68

nin

-0.68

loading

-0.67

cffffcc

-0.66

POSITIVE LOGITS

soever

0.92

 Makes

0.83

 Lies

0.83

 distinguishes

0.79

 separates

0.78

 Choose

0.75

 Definitions

0.74

 Emails

0.73

 Facts

0.73

 Changes

0.71

Activations Density 0.091%

questions and commands

questions and inquiries starting with "What" or "Why."

No Comments

No Known Activations