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

 Wall

-0.66

 Cantor

-0.64

 Sabb

-0.63

 gorilla

-0.62

 damp

-0.60

 Bash

-0.58

 Blow

-0.57

 Santorum

-0.56

Zip

-0.56

 Bond

-0.55

POSITIVE LOGITS

encer

0.86

atform

0.80

acio

0.74

odcast

0.73

gg

0.73

acements

0.71

encers

0.70

abase

0.70

az

0.70

aters

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.