INDEX

Explanations

dates or years mentioned in a specific format within a context

oai_token-act-pair · gpt-3.5-turbo

mentions of historical publications and influential academic works

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

NRL

-0.53

 tyre

-0.53

 outage

-0.52

 Bulldogs

-0.51

 touchdown

-0.51

 tyres

-0.51

 burgers

-0.51

 IMAGES

-0.49

 horns

-0.49

 restart

-0.48

POSITIVE LOGITS

.—

1.24

.*

1.19

.</

1.16

*.

1.16

.[

1.15

.(

1.08

.ï¿½

1.08

.]

1.07

_.

1.06

.<

1.04

Activations Density 0.873%

dates or years mentioned in a specific format within a context

mentions of historical publications and influential academic works

No Comments

No Known Activations