INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Ruins

-0.74

 Riot

-0.71

 Emanuel

-0.70

odes

-0.68

 Shack

-0.65

 Binding

-0.63

à¦

-0.60

WS

-0.58

 Dragonbound

-0.58

 Defenders

-0.58

POSITIVE LOGITS

 cancer

0.86

ansas

0.82

iamond

0.71

cand

0.69

 stakes

0.68

igible

0.67

rounded

0.66

kai

0.66

 leukemia

0.66

 cancers

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.