INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

phrases indicating geographical locations or jurisdictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

VALUE

-0.76

 inval

-0.70

-0.68

Quantity

-0.67

Features

-0.67

ãĢĤ

-0.66

 sensations

-0.66

Ingredients

-0.66

èĢ

-0.65

 entit

-0.65

POSITIVE LOGITS

raged

0.72

wrote

0.71

stood

0.70

hran

0.70

ongyang

0.69

pires

0.69

fecture

0.68

 Concord

0.67

 particular

0.66

 Glou

0.65

Activations Density 0.198%

locations or places

phrases indicating geographical locations or jurisdictions

No Comments

No Known Activations