INDEX

Explanations

research-related terms or phrases

oai_token-act-pair · gpt-3.5-turbo

mentions of research institutions and projects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 femin

-0.66

 Garfield

-0.64

âĶĢâĶĢ

-0.62

 cringe

-0.61

 Rowling

-0.61

anza

-0.58

 compress

-0.58

icago

-0.57

 cracks

-0.57

olson

-0.57

POSITIVE LOGITS

 Laboratories

0.95

 Laboratory

0.94

 Institute

0.85

 Associates

0.85

 Research

0.81

Research

0.81

 Researchers

0.80

 Scientist

0.79

 Ethics

0.79

 Center

0.78

Activations Density 0.022%

research-related terms or phrases

mentions of research institutions and projects

No Comments

No Known Activations