INDEX

Explanations

references to things that are freely available or accessible

oai_token-act-pair · gpt-3.5-turbo

references to the concept of freedom or unrestricted availability

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

senal

-0.76

 Ancients

-0.71

 Stories

-0.70

ochond

-0.69

retty

-0.68

ents

-0.67

arya

-0.67

 Reaction

-0.66

arella

-0.66

 Derby

-0.66

POSITIVE LOGITS

 unrestricted

0.83

 freely

0.82

 flowing

0.79

 roam

0.78

 distribut

0.77

 distribute

0.75

zing

0.73

 unchecked

0.72

bies

0.71

hibited

0.70

Activations Density 0.013%

references to things that are freely available or accessible

references to the concept of freedom or unrestricted availability

No Comments

No Known Activations