INDEX

Explanations

phrases related to causing harm to oneself or others

oai_token-act-pair · gpt-3.5-turbo

references to death and killing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

soType

-0.83

eret

-0.82

taboola

-0.76

Wide

-0.76

soDeliveryDate

-0.73

å§«

-0.72

ĸļ

-0.70

tv

-0.69

URI

-0.68

worthiness

-0.67

POSITIVE LOGITS

 innocent

0.84

 unborn

0.81

 unarmed

0.81

 intruder

0.80

 terrorists

0.78

 messenger

0.77

 senseless

0.75

 classmate

0.74

 murderer

0.74

 crap

0.73

Activations Density 0.161%

phrases related to causing harm to oneself or others

references to death and killing

No Comments

No Known Activations