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

kamp

-0.75

ivals

-0.71

 blitz

-0.64

Charges

-0.64

onz

-0.63

 Jarrett

-0.63

 Teddy

-0.61

 abstinence

-0.60

Jab

-0.60

 Frazier

-0.59

POSITIVE LOGITS

 destro

0.65

english

0.65

oresc

0.61

orne

0.61

arna

0.61

 widened

0.61

herry

0.61

qua

0.60

orno

0.60

currently

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.