INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Thou

-0.71

 Ragnarok

-0.70

 Islands

-0.70

 Samoa

-0.69

 apples

-0.66

 Majesty

-0.65

isconsin

-0.64

Tir

-0.64

rifice

-0.62

 thou

-0.61

POSITIVE LOGITS

ä½ľ

0.75

Entry

0.71

iology

0.69

NetMessage

0.68

 emer

0.68

plays

0.66

RAW

0.65

uses

0.65

onz

0.64

nec

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.