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

ymm

-0.74

tg

-0.66

eu

-0.65

cro

-0.64

flat

-0.62

oji

-0.61

ucl

-0.60

hare

-0.59

ipt

-0.58

apy

-0.58

POSITIVE LOGITS

 Cosponsors

0.96

PDATE

0.79

ONSORED

0.75

resso

0.73

Reviewer

0.71

 Shroud

0.70

Answer

0.68

SAY

0.67

ASS

0.67

ACTED

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.