INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãĤ´ãĥ³

-0.80

Ãį

-0.74

çļ

-0.74

roots

-0.74

ictionary

-0.71

 masc

-0.71

tal

-0.71

å£«

-0.70

owers

-0.70

dan

-0.70

POSITIVE LOGITS

 perjury

0.72

 solicitation

0.70

 completion

0.69

 receipt

0.69

 Vegas

0.68

 Compton

0.64

 Houth

0.62

 ransomware

0.62

 precursor

0.62

qus

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.