INDEX

Explanations

instances of words related to controls, regulations, or instructions

oai_token-act-pair · gpt-3.5-turbo

references to regulatory measures or restrictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Apostle

-0.76

à¤

-0.72

obo

-0.70

ographies

-0.69

 Sacrament

-0.68

union

-0.68

Else

-0.67

nee

-0.66

Asia

-0.66

 apostle

-0.65

POSITIVE LOGITS

 controls

1.03

 controlling

0.89

creen

0.82

vier

0.81

ettings

0.79

ulators

0.76

 Controls

0.74

hops

0.72

 governing

0.72

 guarding

0.71

Activations Density 0.014%

instances of words related to controls, regulations, or instructions

references to regulatory measures or restrictions

No Comments

No Known Activations