INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

lua

-0.83

storage

-0.74

acious

-0.70

 Trials

-0.70

Storage

-0.68

VID

-0.68

inas

-0.67

Tes

-0.66

ials

-0.66

ulkan

-0.66

POSITIVE LOGITS

 tender

0.69

leeve

0.65

 dissolved

0.64

pac

0.64

ence

0.64

yang

0.62

IE

0.62

moderate

0.61

 mutated

0.61

 abolished

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.