INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Mechanics

-0.66

 Females

-0.65

 Merchants

-0.64

lington

-0.64

 Veter

-0.63

Own

-0.62

 Pont

-0.61

 prostitution

-0.60

dt

-0.59

 Maxim

-0.59

POSITIVE LOGITS

ergic

0.77

OWS

0.74

oiler

0.73

OIL

0.66

REAM

0.65

leground

0.65

BOOK

0.64

OUT

0.64

AG

0.64

OULD

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.