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

ãĥī

-0.82

Ãª

-0.72

Magikarp

-0.72

wagen

-0.72

ï

-0.67

 Kinnikuman

-0.61

 Karn

-0.61

.—

-0.60

amac

-0.59

ridor

-0.58

POSITIVE LOGITS

illin

0.87

ovie

0.79

isites

0.77

actionDate

0.74

Jews

0.71

uckland

0.69

iants

0.69

INST

0.69

itory

0.68

codes

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.