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

OUP

-0.94

 perspect

-0.77

 therap

-0.76

ebus

-0.70

 shenan

-0.67

IGHTS

-0.62

 dred

-0.61

 prelim

-0.59

MIS

-0.58

usting

-0.58

POSITIVE LOGITS

0.78

itcher

0.73

tis

0.69

enhagen

0.69

lest

0.66

ordinate

0.65

anus

0.64

atal

0.63

\">

0.63

tar

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.