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

atters

-0.78

 smokes

-0.76

 reproduction

-0.76

 batter

-0.70

 waters

-0.70

 scouts

-0.69

Ń·

-0.67

 eclipse

-0.67

 singles

-0.66

 stacked

-0.65

POSITIVE LOGITS

quickShipAvailable

0.80

ign

0.75

ãĥŁ

0.72

by

0.72

lect

0.72

fare

0.72

norm

0.70

parent

0.70

Closure

0.70

wright

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.