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

reth

-0.71

 fellows

-0.70

aceutical

-0.68

 manned

-0.66

 COUR

-0.66

verages

-0.66

¬¼

-0.65

amera

-0.65

 Runs

-0.64

projects

-0.64

POSITIVE LOGITS

 deposition

0.70

 Tone

0.69

izu

0.68

Tell

0.68

deen

0.67

Nit

0.65

nit

0.64

 smoking

0.64

leigh

0.63

clave

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.