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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 salmon

-0.67

 skilled

-0.67

 caution

-0.66

 tentative

-0.63

 humidity

-0.63

 nearest

-0.61

 Salmon

-0.61

spection

-0.61

 grazing

-0.61

vet

-0.60

POSITIVE LOGITS

re

1.34

adan

0.82

mson

0.80

?]

0.77

adj

0.77

arat

0.77

rez

0.77

advertising

0.76

DeL

0.75

Kre

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.