INDEX

Explanations

quotations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 scheduled

-0.77

 developmental

-0.73

 drown

-0.67

 rall

-0.67

 bashing

-0.65

 sympath

-0.64

 grasp

-0.64

 periodic

-0.63

 derby

-0.62

 taxp

-0.62

POSITIVE LOGITS

We

1.02

1.00

evidence

0.97

no

0.97

Hey

0.92

Dear

0.92

They

0.91

Never

0.90

Technology

0.89

they

0.87

Activations Density 0.208%

quotations

No Comments

No Known Activations