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

BOOK

-0.89

photo

-0.75

 Catal

-0.68

cler

-0.67

psy

-0.65

 Chapman

-0.65

aur

-0.65

ery

-0.63

mon

-0.61

PsyNetMessage

-0.59

POSITIVE LOGITS

aundering

0.74

 Oaks

0.73

ledged

0.72

worms

0.72

culosis

0.72

 ramps

0.71

TPS

0.71

 tiss

0.69

isites

0.68

cellence

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.