INDEX

Explanations

words related to scams or deceptive practices

oai_token-act-pair · gpt-3.5-turbo

words related to deceptive or dubious actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 craving

-0.74

 lapse

-0.71

 reckoning

-0.64

 unsett

-0.64

 landmark

-0.63

 reliance

-0.63

 resurg

-0.61

 mandatory

-0.61

 overdue

-0.61

 bloc

-0.60

POSITIVE LOGITS

cheon

1.01

lator

0.85

 Flavoring

0.85

DAQ

0.78

natureconservancy

0.75

inary

0.75

udo

0.74

esy

0.74

ById

0.74

itiveness

0.74

Activations Density 0.122%

words related to scams or deceptive practices

words related to deceptive or dubious actions

No Comments

No Known Activations