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

FAQ

-0.71

Reilly

-0.68

bb

-0.66

ea

-0.66

BST

-0.65

 Bild

-0.65

orship

-0.64

ODE

-0.64

 Cricket

-0.64

john

-0.64

POSITIVE LOGITS

pointers

0.89

ĺħ

0.78

 desks

0.78

 oneself

0.74

soever

0.71

products

0.71

 wrists

0.66

 brace

0.66

units

0.65

resource

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.