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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 attest

-0.77

 testify

-0.70

 depletion

-0.69

 Papers

-0.66

tif

-0.66

 disapproval

-0.64

 revolt

-0.64

 contag

-0.63

FTWARE

-0.63

 questioning

-0.63

POSITIVE LOGITS

Score

0.76

ħ

0.74

Length

0.72

dn

0.71

Elect

0.69

eous

0.68

 Emin

0.67

Posted

0.67

Redd

0.67

chool

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.