INDEX

Explanations

mentions of job-related terms

oai_token-act-pair · gpt-3.5-turbo

references to employment and job-related themes

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

 carriers

-0.76

 floats

-0.68

 carrier

-0.68

 conceal

-0.67

 conf

-0.65

 discs

-0.64

ribut

-0.64

 protocols

-0.61

itol

-0.61

olis

-0.61

POSITIVE LOGITS

Job

4.10

Job

2.89

job

2.24

 Jobs

1.96

jobs

1.92

job

1.78

 jobs

1.46

 Employment

1.31

 Occupations

1.28

 Worker

1.22

Activations Density 0.022%

mentions of job-related terms

references to employment and job-related themes

No Comments

No Known Activations