INDEX

Explanations

profession-related terms, such as job titles, industries, and areas of expertise

oai_token-act-pair · gpt-3.5-turbo

references to various professional roles and stakeholders involved in collaborative projects or discussions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Jericho

-0.80

 Strawberry

-0.73

 Rampage

-0.73

 Lethal

-0.72

 Everest

-0.71

 Invasion

-0.69

 upside

-0.69

 HELL

-0.69

 Passage

-0.69

 Conversation

-0.68

POSITIVE LOGITS

 alike

1.47

 professionals

1.06

 allies

1.04

 educators

1.02

 enthusiasts

0.99

 agencies

0.98

 experts

0.98

 contractors

0.97

 clergy

0.97

 specialists

0.96

Activations Density 0.374%

profession-related terms, such as job titles, industries, and areas of expertise

references to various professional roles and stakeholders involved in collaborative projects or discussions

No Comments

No Known Activations