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

acet

-0.71

gm

-0.69

 targ

-0.65

 sund

-0.65

ingu

-0.62

ppa

-0.62

 artificially

-0.62

PST

-0.62

alli

-0.62

 honors

-0.62

POSITIVE LOGITS

DERR

0.95

continue

0.70

Graph

0.69

ndra

0.68

 Brookings

0.68

Ze

0.67

STEP

0.67

went

0.66

 WATCHED

0.66

ItemImage

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.