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

 Mysteries

-0.71

-0.67

Magikarp

-0.66

opsy

-0.65

 Institutes

-0.64

 Ernst

-0.63

 Prediction

-0.63

 Indo

-0.62

pmwiki

-0.62

ypes

-0.62

POSITIVE LOGITS

ively

0.71

evin

0.69

mods

0.68

 blockers

0.65

leasing

0.63

gui

0.62

ilver

0.62

ou

0.61

MTA

0.61

plugins

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.