INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ult

-0.79

requires

-0.78

onia

-0.74

imates

-0.74

ames

-0.73

val

-0.73

opted

-0.72

paces

-0.72

umerable

-0.70

ensitivity

-0.69

POSITIVE LOGITS

 Shoot

0.64

 batteries

0.64

Alz

0.64

 Arrest

0.63

 Rookie

0.62

 Levy

0.62

 Shooter

0.61

RI

0.59

 refriger

0.59

 plaque

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.