INDEX

Explanations

phrases related to actions accompanied by consequences or reactions

oai_token-act-pair · gpt-3.5-turbo

instances of reported actions or statements from the subject about threats and requests

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

llah

-0.67

cation

-0.65

isal

-0.64

etheless

-0.63

mania

-0.62

aml

-0.62

panic

-0.62

illo

-0.62

culus

-0.61

lance

-0.58

POSITIVE LOGITS

 themselves

1.07

 selves

0.89

selves

0.77

MpServer

0.68

 helmets

0.67

 mouths

0.66

 uniforms

0.64

 microphones

0.60

 orbits

0.60

 jointly

0.60

Activations Density 0.834%

phrases related to actions accompanied by consequences or reactions

instances of reported actions or statements from the subject about threats and requests

No Comments

No Known Activations