INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ATS

-0.82

allah

-0.74

HR

-0.73

 COMMUN

-0.72

owitz

-0.71

network

-0.71

ALTH

-0.70

emet

-0.70

ythm

-0.69

ROR

-0.68

POSITIVE LOGITS

 paran

0.82

 Seym

0.72

 frivol

0.70

 pard

0.68

 rebellious

0.67

 spons

0.66

 defiant

0.65

 juven

0.65

 besie

0.64

 challeng

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.