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

Interstitial

-0.74

FML

-0.73

phasis

-0.72

enment

-0.68

amines

-0.67

 GOODMAN

-0.65

DEV

-0.65

tle

-0.64

hyde

-0.63

heid

-0.62

POSITIVE LOGITS

unic

0.69

esian

0.68

 virgin

0.66

oi

0.65

fen

0.62

nis

0.62

Republic

0.61

 skiing

0.59

gress

0.58

 frig

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.