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

EStreamFrame

-0.80

RH

-0.73

pl

-0.72

paragraph

-0.71

RP

-0.71

psc

-0.69

platform

-0.68

pr

-0.68

glas

-0.68

people

-0.68

POSITIVE LOGITS

 Virus

0.76

 Tanks

0.70

 Thumbnails

0.67

 Quant

0.65

Isa

0.64

 Tests

0.62

 Imran

0.59

 Gors

0.59

aters

0.59

 Shame

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.