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

pedia

-0.77

 Krishna

-0.72

 motel

-0.69

ibrary

-0.68

potion

-0.66

PsyNetMessage

-0.63

eria

-0.62

 Marketable

-0.61

ensen

-0.60

ages

-0.60

POSITIVE LOGITS

æĸ¹

0.68

udo

0.68

nesday

0.65

dB

0.65

è¿

0.64

aband

0.64

onic

0.62

dB

0.62

Jed

0.62

raltar

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.