INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

reader

-0.72

 advice

-0.68

 deposition

-0.64

 anonymity

-0.62

ials

-0.61

nsics

-0.59

ay

-0.58

 heck

-0.58

 shading

-0.58

tips

-0.58

POSITIVE LOGITS

 Nept

0.75

anty

0.70

Interstitial

0.70

pta

0.69

ione

0.69

ilitary

0.68

 Kard

0.67

mented

0.65

bley

0.65

spir

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.