INDEX

Explanations

names of medical institutions or disciplines

oai_token-act-pair · gpt-3.5-turbo

references to medical journals and institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

opp

-0.68

agg

-0.67

bon

-0.64

gb

-0.64

fin

-0.64

gr

-0.63

 oblig

-0.62

 cruc

-0.61

 offline

-0.61

grab

-0.60

POSITIVE LOGITS

 Medicine

3.95

 medicine

2.20

Medic

1.69

 Surgery

1.57

 Physicians

1.49

 Psychiatry

1.45

Doctors

1.43

 Medical

1.43

 Diseases

1.38

 Pharma

1.37

Activations Density 0.010%

names of medical institutions or disciplines

references to medical journals and institutions

No Comments

No Known Activations