INDEX

Explanations

expressions related to penetration, whether physical or metaphorical

oai_token-act-pair · gpt-3.5-turbo

terms related to penetration or entry

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

HCR

-0.74

umble

-0.72

makers

-0.68

uf

-0.63

inn

-0.63

maker

-0.61

arb

-0.61

çĶŁ

-0.60

corn

-0.60

ÄŁ

-0.59

POSITIVE LOGITS

 penetrated

1.24

 penetration

1.22

 penetrate

1.20

 penet

1.11

 penetrating

0.99

 Penet

0.81

atform

0.81

 deeper

0.79

 Explorer

0.77

 embr

0.75

Activations Density 0.024%

expressions related to penetration, whether physical or metaphorical

terms related to penetration or entry

No Comments

No Known Activations