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

bane

-0.82

jah

-0.74

keley

-0.72

rawdownloadcloneembedreportprint

-0.69

Cause

-0.67

Jesus

-0.65

DERR

-0.65

xff

-0.65

shaw

-0.65

jab

-0.62

POSITIVE LOGITS

pread

0.74

 dipping

0.71

emi

0.67

arching

0.65

eatures

0.65

 concess

0.65

 operating

0.65

 servicing

0.63

inia

0.63

vertisement

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.