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

abal

-0.80

 claimant

-0.68

washer

-0.67

 displacement

-0.66

Smy

-0.63

 justice

-0.61

 defence

-0.60

 particulars

-0.60

 occupant

-0.60

anes

-0.59

POSITIVE LOGITS

ioxide

0.86

ãģķ

0.86

ï¸ı

0.82

ãĥ©ãĥ³

0.80

natureconservancy

0.75

ãĥŃ

0.73

rh

0.73

ÙĨ

0.73

Revolution

0.73

pora

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.