INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Destination

-0.71

 targ

-0.65

 prompt

-0.65

 spark

-0.64

 hostages

-0.63

uers

-0.62

 staging

-0.62

 aggress

-0.61

 creditors

-0.61

ups

-0.60

POSITIVE LOGITS

thia

0.89

hedral

0.81

ãĥ´ãĤ¡

0.80

âĸ¬âĸ¬

0.78

thur

0.74

stros

0.74

hee

0.73

SHIP

0.73

gdala

0.73

IRE

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.