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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

uminati

-0.90

 Peach

-0.69

noon

-0.67

 Nost

-0.64

iley

-0.63

liest

-0.63

 Celestial

-0.63

 studios

-0.62

 Patreon

-0.62

 reapp

-0.61

POSITIVE LOGITS

zsche

0.94

jong

0.93

theless

0.84

 planner

0.71

 agre

0.71

agic

0.67

Handler

0.67

Beh

0.66

erker

0.66

ebus

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.