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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ancest

-0.68

entric

-0.67

cients

-0.67

 ward

-0.62

FB

-0.61

LB

-0.60

 killers

-0.60

oused

-0.60

oming

-0.59

 besie

-0.59

POSITIVE LOGITS

FIELD

0.86

ãĥ¥

0.78

Pont

0.70

grain

0.66

 Echoes

0.65

eca

0.65

esson

0.64

ema

0.62

 Lumpur

0.62

wagen

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.