INDEX

Explanations

No Explanations Found

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 warr

-0.74

 obser

-0.69

 charact

-0.66

]).

-0.66

XT

-0.66

staff

-0.66

gren

-0.66

 asses

-0.64

liest

-0.64

Reviewed

-0.62

POSITIVE LOGITS

addin

0.71

à¨

0.65

 Blues

0.64

 Torch

0.63

 à¨

0.62

 subur

0.62

Pic

0.60

Owl

0.60

vell

0.60

ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.