INDEX

Explanations

words related to regulations, guidelines, or formal procedures

oai_token-act-pair · gpt-3.5-turbo

connections to policies and procedures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

usercontent

-0.86

quished

-0.84

imbabwe

-0.77

ailable

-0.76

uesday

-0.74

utan

-0.72

unic

-0.71

atinum

-0.71

enaries

-0.70

leted

-0.70

POSITIVE LOGITS

 morals

1.27

 philosophies

1.25

 practices

1.22

 habits

1.20

 procedures

1.17

 tactics

1.17

 beliefs

1.15

 attitudes

1.14

 traditions

1.14

 principles

1.13

Activations Density 0.356%

words related to regulations, guidelines, or formal procedures

connections to policies and procedures

No Comments

No Known Activations