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

Mex

-0.74

usted

-0.70

 brushes

-0.66

 Cook

-0.63

UB

-0.63

 corros

-0.62

aska

-0.61

 Khan

-0.60

uda

-0.60

Alb

-0.60

POSITIVE LOGITS

çīĪ

0.83

awa

0.81

phasis

0.75

izen

0.72

oso

0.70

srf

0.69

wake

0.68

ticket

0.68

visory

0.67

Oo

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.