INDEX

Explanations

words related to a specific spelling or form (orm) pattern

oai_token-act-pair · gpt-3.5-turbo

terms related to forms and structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Decker

-0.70

prints

-0.64

Fey

-0.64

 Rico

-0.63

 Absent

-0.62

 Maya

-0.62

print

-0.60

cho

-0.60

 Marcos

-0.59

Favorite

-0.58

POSITIVE LOGITS

orph

1.16

ont

1.06

orm

1.04

orthy

1.00

otive

0.96

inational

0.93

etheus

0.93

endez

0.93

ajor

0.92

yss

0.91

Activations Density 0.011%

words related to a specific spelling or form (orm) pattern

terms related to forms and structures

No Comments

No Known Activations