INDEX

Explanations

references to technology interfaces or digital user experiences

oai_token-act-pair · gpt-3.5-turbo

references to the overall quality or nature of various experiences

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

laws

-0.76

fam

-0.71

law

-0.70

nda

-0.68

 subdiv

-0.65

roup

-0.64

 stockp

-0.63

pta

-0.62

sub

-0.61

 filings

-0.61

POSITIVE LOGITS

 Experience

0.89

Experience

0.88

IENCE

0.87

ULTS

0.86

ually

0.84

 Royale

0.81

HAEL

0.80

iences

0.80

 Remastered

0.78

 experien

0.78

Activations Density 0.034%

references to technology interfaces or digital user experiences

references to the overall quality or nature of various experiences

No Comments

No Known Activations