INDEX

Explanations

phrases mentioning specific institutions or organizations

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in" used in various contexts

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

DragonMagazine

-0.85

 gaping

-0.73

0000000000000000

-0.71

 fortun

-0.70

untled

-0.69

 smugglers

-0.69

*/

-0.68

ĺħ

-0.68

poll

-0.68

 innoc

-0.68

POSITIVE LOGITS

 conjunction

1.20

 collaboration

1.11

 partnership

1.11

 Pasadena

1.08

 Washington

1.07

 Philadelphia

1.06

Los

1.01

 Melbourne

1.01

 California

0.98

 London

0.98

Activations Density 0.106%

phrases mentioning specific institutions or organizations

instances of the word "in" used in various contexts

No Comments

No Known Activations