INDEX

Explanations

phrases or words indicating a false justification or reason for a particular action or belief

oai_token-act-pair · gpt-3.5-turbo

terms related to justifications or excuses for actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

elong

-0.74

Life

-0.70

itivity

-0.70

Die

-0.69

 evolve

-0.68

igr

-0.63

AMI

-0.63

average

-0.62

Surv

-0.62

life

-0.62

POSITIVE LOGITS

 pretext

3.84

 guise

2.09

 spurious

1.38

 bogus

1.36

 provocation

1.29

 disguise

1.24

 phony

1.24

 dubious

1.22

 pretended

1.18

 euphem

1.18

Activations Density 0.027%

phrases or words indicating a false justification or reason for a particular action or belief

terms related to justifications or excuses for actions

No Comments

No Known Activations