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

TT

-0.70

Pwr

-0.70

hus

-0.70

morning

-0.69

bage

-0.67

TW

-0.67

uties

-0.66

Roberts

-0.65

lict

-0.65

utterstock

-0.65

POSITIVE LOGITS

RSA

0.69

 Ether

0.64

Obj

0.62

 offered

0.60

 indu

0.60

 Frie

0.59

jon

0.59

 Contracts

0.59

RFC

0.58

 FranÃ§

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.