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

intendent

-0.86

ldon

-0.78

tics

-0.73

 Redux

-0.73

uggest

-0.71

venient

-0.71

arcity

-0.70

rency

-0.70

rals

-0.69

 mathemat

-0.68

POSITIVE LOGITS

advertisement

0.72

 genitals

0.65

Fax

0.62

Cause

0.62

 anus

0.61

GROUP

0.61

 genital

0.60

 channel

0.59

 Channel

0.59

-->

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.