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

OOK

-0.66

ESE

-0.65

 Shining

-0.62

Browser

-0.62

ĻĤ

-0.62

atically

-0.61

Ĥ¬

-0.60

onet

-0.60

 beet

-0.60

GF

-0.59

POSITIVE LOGITS

NRS

0.76

mA

0.70

lando

0.69

rast

0.68

HOU

0.65

antage

0.64

 suppl

0.63

iors

0.61

 Stability

0.60

tops

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.