INDEX

Explanations

references to gender differences, specifically focusing on boys

oai_token-act-pair · gpt-3.5-turbo

repeated references to boys

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

itures

-0.79

lyak

-0.77

iture

-0.75

arily

-0.73

 Solitaire

-0.70

acle

-0.70

Las

-0.68

igation

-0.68

office

-0.67

iped

-0.67

POSITIVE LOGITS

 ages

1.02

 boys

1.01

 Scouts

0.99

friend

0.93

 puberty

0.91

volent

0.91

boys

0.90

hood

0.86

 scout

0.81

bands

0.80

Activations Density 0.041%

references to gender differences, specifically focusing on boys

repeated references to boys

No Comments

No Known Activations