INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

vana

-0.72

 Donation

-0.68

 transparency

-0.66

plex

-0.65

 frontline

-0.64

 pree

-0.62

 uniqueness

-0.62

anchester

-0.62

 clarity

-0.62

 begg

-0.61

POSITIVE LOGITS

aukee

0.87

oit

0.87

mare

0.75

 sailed

0.70

ento

0.68

IQ

0.68

Wind

0.66

urnal

0.66

numbered

0.66

 Amend

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.