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

 awaited

-0.68

hunt

-0.68

bard

-0.67

ELD

-0.67

Mos

-0.67

Fal

-0.66

Topic

-0.63

icht

-0.63

 Agent

-0.62

inki

-0.61

POSITIVE LOGITS

ĪĴ

0.77

ignt

0.76

 Olivier

0.76

eneg

0.73

©¶æ

0.73

apon

0.72

eatures

0.69

ographs

0.69

 Seym

0.68

mpeg

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.