INDEX

Explanations

phrases related to legal actions involving individuals

oai_token-act-pair · gpt-3.5-turbo

phrases indicating ongoing or repeated actions, especially those related to crimes or events involving individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Craft

-0.80

Myth

-0.75

 Cavern

-0.74

reality

-0.73

 hegemony

-0.69

 Illusion

-0.67

é¾įå

-0.67

Struct

-0.67

 Agric

-0.66

Universal

-0.66

POSITIVE LOGITS

 arrested

1.01

 arra

0.92

 deceased

0.92

 suspected

0.90

 jailed

0.90

 reportedly

0.89

 injured

0.89

 sentenced

0.89

 hospitalized

0.88

 flown

0.86

Activations Density 0.389%

phrases related to legal actions involving individuals

phrases indicating ongoing or repeated actions, especially those related to crimes or events involving individuals

No Comments

No Known Activations