INDEX

Explanations

titles or descriptions indicating the last installment of a series or a particular edition, usually associated with high intensity or impact

oai_token-act-pair · gpt-3.5-turbo

closing parentheses in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 answ

-0.82

 artif

-0.77

tering

-0.71

Ļ½

-0.68

clus

-0.68

¬¼

-0.67

bably

-0.67

 compan

-0.66

 recru

-0.65

 footing

-0.64

POSITIVE LOGITS

 âĵĺ

0.90

=>

0.83

 âĨĴ

0.79

0.78

↵

0.77

->

0.76

↵↵

0.75

Ltd

0.75

 Committees

0.74

 Races

0.74

Activations Density 0.130%

titles or descriptions indicating the last installment of a series or a particular edition, usually associated with high intensity or impact

closing parentheses in the text

No Comments

No Known Activations