INDEX

Explanations

mentions of drunkenness or drunk-related activities, especially focused on drunk driving

oai_token-act-pair · gpt-3.5-turbo

references to intoxication or impairment due to alcohol

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Flavoring

-0.98

pta

-0.82

pha

-0.76

JPM

-0.73

isite

-0.73

TPP

-0.70

ocol

-0.66

 Hosp

-0.66

CVE

-0.66

Downloadha

-0.65

POSITIVE LOGITS

ards

0.95

 manslaughter

0.89

 drunk

0.86

 driving

0.84

driving

0.84

ard

0.83

 Driving

0.79

 underage

0.78

 arrest

0.76

bott

0.74

Activations Density 0.064%

mentions of drunkenness or drunk-related activities, especially focused on drunk driving

references to intoxication or impairment due to alcohol

No Comments

No Known Activations