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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

meyer

-0.66

 enthus

-0.65

EMA

-0.64

 latent

-0.63

iris

-0.63

 overflowing

-0.62

azy

-0.61

 inert

-0.60

 cannabin

-0.60

 clos

-0.59

POSITIVE LOGITS

stones

0.72

aires

0.68

ranch

0.66

emn

0.65

holes

0.64

ify

0.62

acles

0.62

 Colo

0.61

obl

0.61

 Devin

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.