INDEX

Explanations

proper names related to people or places

oai_token-act-pair · gpt-3.5-turbo

proper names of people and organizations

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

ocard

-0.87

ancers

-0.80

visor

-0.79

ikers

-0.78

psons

-0.74

Score

-0.74

onder

-0.72

arm

-0.72

ancer

-0.71

 Antar

-0.69

POSITIVE LOGITS

 Barnett

1.33

Barn

0.71

 McMahon

0.69

 Abbott

0.69

 Ingram

0.69

TAIN

0.69

 McGee

0.67

Pax

0.67

 mandates

0.67

rules

0.66

Activations Density 0.017%

proper names related to people or places

proper names of people and organizations

No Comments

No Known Activations