INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 inspir

-0.66

Fut

-0.66

vich

-0.64

 Siber

-0.64

 freaking

-0.63

 sequels

-0.62

lif

-0.61

 updates

-0.61

 Alien

-0.59

ondon

-0.58

POSITIVE LOGITS

omen

0.82

qu

0.76

aders

0.76

ector

0.75

 Pearson

0.73

cot

0.73

igma

0.73

omer

0.72

displayText

0.71

opes

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.