INDEX

Explanations

various technical terms and measurements, including abbreviations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 princ

-0.80

 accrued

-0.74

 trump

-0.70

 Samar

-0.70

 cher

-0.69

 stewards

-0.69

 holidays

-0.65

 parasite

-0.65

 imperson

-0.64

 Respect

-0.64

POSITIVE LOGITS

MF

1.77

CB

1.75

RW

1.74

LB

1.74

BG

1.74

CM

1.74

DK

1.72

1.71

Activations Density 0.101%

various technical terms and measurements, including abbreviations

No Comments

No Known Activations