INDEX

Explanations

words related to houses or households

oai_token-act-pair · gpt-3.5-turbo

multiple references to the concept of homes

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Delta

-0.77

iasis

-0.69

BILITY

-0.67

 Gaul

-0.65

 guiActiveUn

-0.63

 Lance

-0.61

++++++++++++++++

-0.60

ional

-0.60

 Period

-0.59

 Adinida

-0.59

POSITIVE LOGITS

chool

1.60

pun

0.99

stead

0.92

pring

0.91

creen

0.89

 dwelling

0.82

 dwellings

0.78

ports

0.78

 homes

0.77

ourced

0.77

Activations Density 0.015%

words related to houses or households

multiple references to the concept of homes

No Comments

No Known Activations