INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Redd

-0.83

 Thousand

-0.76

Pu

-0.74

 Tong

-0.69

pun

-0.68

Yi

-0.68

 Trin

-0.67

 Origin

-0.67

 Kard

-0.66

 Twins

-0.65

POSITIVE LOGITS

gemony

0.74

phrine

0.69

otomy

0.68

lement

0.68

rael

0.68

gian

0.65

azo

0.65

arte

0.65

arity

0.64

obby

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.