INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

iago

-0.69

 Knox

-0.63

acle

-0.63

rome

-0.62

 daemon

-0.62

 imitation

-0.61

aukee

-0.61

rists

-0.60

 lockout

-0.60

 signature

-0.59

POSITIVE LOGITS

Interstitial

0.80

ãĤ¤ãĥĪ

0.77

 according

0.76

 Rooms

0.73

ccording

0.73

 According

0.72

×Ļ

0.69

partial

0.69

according

0.69

 Fields

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.