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

uart

-0.76

ItemTracker

-0.73

 Palest

-0.72

 Ellison

-0.72

Ô

-0.70

-0.69

APD

-0.68

 sshd

-0.67

=#

-0.66

POSITIVE LOGITS

istries

0.76

gary

0.66

 prob

0.64

Gab

0.63

hered

0.63

 juice

0.61

 blocker

0.61

gars

0.60

nect

0.60

 counselor

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.