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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ember

-0.77

nesday

-0.74

aughtered

-0.72

alion

-0.71

emon

-0.70

alos

-0.69

 Bastard

-0.68

resy

-0.67

ir

-0.67

essage

-0.67

POSITIVE LOGITS

¹

0.73

¾

0.73

µ

0.68

Â¶

0.65

Ļ½

0.64

ī

0.63

å§

0.62

 therapy

0.61

scape

0.61

pill

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.