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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

byss

-0.62

Goal

-0.61

 Hull

-0.61

 Costa

-0.61

seys

-0.60

Tokens

-0.60

 estab

-0.59

 cruiser

-0.59

 attitude

-0.58

 Bucc

-0.58

POSITIVE LOGITS

>]

0.75

enza

0.74

icy

0.73

cised

0.71

obyl

0.70

ONY

0.70

avis

0.70

ellen

0.69

CDC

0.68

dust

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.