INDEX

Explanations

negatively perceived words or phrases

oai_token-act-pair · gpt-3.5-turbo

expressions of enjoyment or dissatisfaction in reviews

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ewitness

-0.87

 telephone

-0.73

ospons

-0.72

 Allaah

-0.69

 ..."

-0.69

estern

-0.68

upper

-0.68

Federal

-0.68

fax

-0.67

illegal

-0.66

POSITIVE LOGITS

 nostalgia

0.91

 Whedon

0.89

 pacing

0.89

 fanbase

0.87

 indie

0.84

 tropes

0.84

 cinematic

0.84

 genuinely

0.83

 goddamn

0.83

 Konami

0.81

Activations Density 1.126%

negatively perceived words or phrases

expressions of enjoyment or dissatisfaction in reviews

No Comments

No Known Activations