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

Leaks

-0.75

WP

-0.66

CLOSE

-0.62

nces

-0.61

Motion

-0.60

 disclaimer

-0.59

LY

-0.59

Abstract

-0.59

BIL

-0.58

 Availability

-0.58

POSITIVE LOGITS

rant

1.09

rants

0.82

schild

0.74

rim

0.72

rete

0.71

mast

0.69

 Aram

0.68

rir

0.68

rael

0.67

adia

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.