INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 vein

-0.72

ther

-0.68

 paraph

-0.67

 annotations

-0.66

cture

-0.65

 rebutt

-0.62

ner

-0.62

 stump

-0.61

 kilomet

-0.60

aly

-0.59

POSITIVE LOGITS

 LIMITED

0.80

REL

0.71

Redditor

0.71

Publisher

0.69

Oracle

0.68

Generic

0.68

 Protective

0.68

Availability

0.67

Roll

0.67

Problem

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.