INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

FACE

-0.66

 chalk

-0.64

luster

-0.64

 regimes

-0.60

SN

-0.60

 Trinidad

-0.60

DUP

-0.59

verage

-0.59

rah

-0.57

phy

-0.57

POSITIVE LOGITS

skirts

0.76

orget

0.70

 guiActiveUnfocused

0.66

 ende

0.66

 Tanz

0.65

uten

0.64

awa

0.64

comings

0.64

 Beir

0.64

irlf

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.