INDEX

Explanations

headphone-related terms

oai_token-act-pair · gpt-3.5-turbo

terms related to headphones and audio equipment

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Lann

-0.87

 Wikimedia

-0.78

 Livingston

-0.70

 Hardy

-0.70

erest

-0.69

Lew

-0.68

 Tory

-0.68

 Expend

-0.68

 Bedford

-0.67

 Craw

-0.67

POSITIVE LOGITS

 headphones

1.64

 headphone

1.41

 amplifier

1.14

 impedance

1.03

awei

1.03

obook

1.00

 enthusi

0.99

 conduc

0.97

ear

0.96

 tremend

0.93

Activations Density 0.012%

headphone-related terms

terms related to headphones and audio equipment

No Comments

No Known Activations