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

 CODE

-0.68

 Cust

-0.65

 Doct

-0.65

 Icar

-0.65

fixes

-0.64

ronic

-0.64

FML

-0.62

btn

-0.61

 proc

-0.61

 Const

-0.61

POSITIVE LOGITS

Synopsis

0.72

handed

0.70

unning

0.70

iltration

0.65

indal

0.63

htaking

0.63

hander

0.62

ount

0.62

itud

0.62

ÂłÂłÂł

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.