INDEX

Explanations

phrases related to legal and governmental entities

oai_token-act-pair · gpt-3.5-turbo

references to acronyms related to organizations or acts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

itars

-0.79

 ç¥ŀ

-0.78

taboola

-0.71

Â¯

-0.70

shown

-0.66

 WATCHED

-0.66

quickShipAvailable

-0.65

 Reincarnated

-0.65

âĢ¢âĢ¢âĢ¢âĢ¢

-0.63

fried

-0.63

POSITIVE LOGITS

HS

0.85

 Directorate

0.82

 Commission

0.79

RC

0.78

 Chair

0.78

 Institute

0.76

 Presidency

0.76

 presidency

0.76

 Staff

0.75

onica

0.74

Activations Density 0.170%

phrases related to legal and governmental entities

references to acronyms related to organizations or acts

No Comments

No Known Activations