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

Reviewer

-0.74

CW

-0.72

DA

-0.68

ol

-0.65

mus

-0.65

ESH

-0.63

tam

-0.63

wat

-0.61

fal

-0.61

nit

-0.61

POSITIVE LOGITS

olor

0.80

urdue

0.73

robe

0.70

yll

0.67

 Emin

0.66

oreal

0.65

è£ħ

0.64

urion

0.64

orus

0.64

 Temper

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.