INDEX

Explanations

phrases mentioning specific units or divisions within various organizations

oai_token-act-pair · gpt-3.5-turbo

references to various specialized units or divisions

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

FORE

-0.78

 fathers

-0.69

lez

-0.67

à¼

-0.66

donald

-0.66

christ

-0.65

ozy

-0.65

extreme

-0.64

 influ

-0.64

 allergic

-0.62

POSITIVE LOGITS

 Unit

3.90

 Units

2.61

Unit

2.46

 unit

2.04

 units

1.63

unit

1.59

units

1.32

 Module

1.24

 Squad

1.22

 Suite

1.15

Activations Density 0.016%

phrases mentioning specific units or divisions within various organizations

references to various specialized units or divisions

No Comments

No Known Activations