INDEX

Explanations

the word "Dale" with differing intensities

oai_token-act-pair · gpt-3.5-turbo

references to the name "Dale" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

rontal

-0.83

sit

-0.72

ivities

-0.71

eous

-0.71

Political

-0.67

pol

-0.67

inventoryQuantity

-0.66

Latin

-0.66

ulatory

-0.65

icipated

-0.65

POSITIVE LOGITS

 Dale

1.32

 Carnegie

0.89

 Burnett

0.87

McF

0.82

 Griff

0.82

 Yates

0.80

leigh

0.78

 Moff

0.76

 Driver

0.75

 Scully

0.74

Activations Density 0.005%

the word "Dale" with differing intensities

references to the name "Dale" in various contexts

No Comments

No Known Activations