INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Tanz

-0.70

 Gavin

-0.70

 Herod

-0.69

-0.68

 chast

-0.66

etheus

-0.66

antine

-0.65

Ada

-0.64

Dag

-0.63

 indices

-0.62

POSITIVE LOGITS

said

0.83

ingham

0.79

oval

0.78

ively

0.76

killed

0.72

ennial

0.71

ividual

0.71

ifted

0.70

arro

0.69

aires

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.