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

imates

-0.71

Owl

-0.70

 Pulitzer

-0.67

 Arrows

-0.63

 Torch

-0.62

 Became

-0.61

 Bullets

-0.61

=#

-0.60

 Easy

-0.59

 Quit

-0.59

POSITIVE LOGITS

 spons

0.81

cephal

0.74

merce

0.72

smoking

0.72

secondary

0.70

 Franch

0.70

adequ

0.70

sych

0.68

ĪĴ

0.67

charged

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.