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

 Tape

-0.73

 Competition

-0.67

Cay

-0.67

 Continued

-0.67

 Glob

-0.66

 Cutter

-0.65

 Cull

-0.64

Fas

-0.63

 Curve

-0.62

Ac

-0.62

POSITIVE LOGITS

ieri

0.82

uesday

0.76

beer

0.73

ocre

0.70

 perjury

0.70

ugu

0.68

berman

0.67

ħĭ

0.67

icus

0.66

mith

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.