INDEX

Explanations

references to specific numerical values, measurements, and units

oai_token-act-pair · gpt-3.5-turbo

titles or names of specific works and reference their contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

href

-0.77

 constitu

-0.76

vim

-0.75

manuel

-0.72

borgh

-0.71

 Mast

-0.71

viron

-0.71

Language

-0.70

Jas

-0.70

BuyableInstoreAndOnline

-0.69

POSITIVE LOGITS

 Round

2.02

 round

1.92

 Square

1.90

 square

1.86

Square

1.83

Round

1.80

round

1.64

square

1.60

 squares

1.53

 oval

1.46

Activations Density 0.192%

references to specific numerical values, measurements, and units

titles or names of specific works and reference their contexts

No Comments

No Known Activations