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

essen

-0.74

SPONSORED

-0.68

 guiActiveUnfocused

-0.64

demand

-0.63

 sensibilities

-0.61

 PLoS

-0.61

 Cult

-0.61

imble

-0.61

 Trend

-0.61

pmwiki

-0.60

POSITIVE LOGITS

uria

0.81

 Flake

0.73

RH

0.73

NL

0.70

 Haley

0.67

 miscar

0.67

 Kang

0.66

 Burnett

0.65

Mew

0.64

backer

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.