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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

metadata

-0.77

etheless

-0.77

DragonMagazine

-0.72

netflix

-0.68

suits

-0.67

ruff

-0.65

ortmund

-0.65

bidden

-0.65

ugi

-0.65

imens

-0.65

POSITIVE LOGITS

 privileged

0.75

ufact

0.68

-+

0.65

Īè

0.65

 grateful

0.64

 occupational

0.62

ensional

0.62

onto

0.61

aton

0.61

 pursuing

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.