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

 creator

-0.71

 creators

-0.70

 spokesperson

-0.66

 brid

-0.65

 Bangl

-0.65

 Himself

-0.64

rawdownloadcloneembedreportprint

-0.63

 Galile

-0.62

 Creator

-0.62

 surve

-0.62

POSITIVE LOGITS

onz

0.80

ombo

0.79

orie

0.78

oday

0.74

fired

0.74

orb

0.72

Operation

0.72

orbit

0.71

oval

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.