INDEX

Explanations

company names that end with "Inc" followed by a high activation value

oai_token-act-pair · gpt-3.5-turbo

mentions of corporate entities or organizations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rame

-0.70

DH

-0.69

wall

-0.67

 drawer

-0.66

jad

-0.65

 Rory

-0.65

wen

-0.64

 answ

-0.62

=>

-0.62

kered

-0.61

POSITIVE LOGITS

Inc

3.71

INC

2.16

LLC

2.13

Inc

2.02

Ltd

1.99

 Corp

1.64

INC

1.49

inc

1.42

LLP

1.38

inc

1.38

Activations Density 0.009%

company names that end with "Inc" followed by a high activation value

mentions of corporate entities or organizations

No Comments

No Known Activations