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

azo

-0.77

pes

-0.77

vantage

-0.77

flix

-0.75

 breathe

-0.72

aved

-0.71

abase

-0.69

seek

-0.69

iddler

-0.67

enos

-0.66

POSITIVE LOGITS

May

1.12

 April

1.07

 June

1.07

 August

1.06

 July

1.04

 February

1.03

 November

1.00

 October

0.98

 January

0.97

 September

0.95

Activations Density 0.000%

No Known Activations

This feature has no known activations.