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

 Grizzlies

-0.76

iled

-0.72

SPONSORED

-0.71

 throats

-0.70

frey

-0.69

ousy

-0.69

iles

-0.68

ials

-0.66

imeo

-0.65

aturday

-0.65

POSITIVE LOGITS

len

0.68

Ern

0.66

 Sinn

0.65

mA

0.64

 Password

0.64

Lau

0.62

Sai

0.62

 Skywalker

0.61

©¶æ¥µ

0.61

MEP

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.