INDEX

Explanations

The neuron is looking for mentions of specific software apps or platforms

oai_token-act-pair · gpt-3.5-turbo

references to specific software applications and features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

manship

-0.83

heses

-0.73

nil

-0.72

jah

-0.71

Tier

-0.68

blow

-0.67

STAR

-0.67

lam

-0.65

Writ

-0.63

loe

-0.63

POSITIVE LOGITS

app

1.20

 latest

1.01

 Settings

1.00

 newest

0.99

 dashboard

0.93

 browser

0.93

 installer

0.92

 default

0.92

 touchscreen

0.91

 emulator

0.90

Activations Density 0.209%

The neuron is looking for mentions of specific software apps or platforms

references to specific software applications and features

No Comments

No Known Activations