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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Objective

-0.68

 Agric

-0.68

 Procedure

-0.67

 premises

-0.65

 DRAGON

-0.63

 swings

-0.62

 slopes

-0.61

 footh

-0.60

 newcomers

-0.60

 Prospect

-0.59

POSITIVE LOGITS

creen

0.91

////

0.83

netflix

0.82

Ĥª

0.76

irst

0.74

////////////////////////////////

0.74

ilver

0.72

hid

0.71

ulk

0.70

////////////////

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.