INDEX

Explanations

mentions or descriptions of food items and dining experiences

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Journals

-0.82

 Gamergate

-0.77

ayson

-0.75

 Bonds

-0.75

 Vance

-0.70

Privacy

-0.69

ipolar

-0.69

Detect

-0.68

etheus

-0.68

 Dynamics

-0.68

POSITIVE LOGITS

 grilled

1.78

 fried

1.70

 burgers

1.67

 salads

1.64

 tacos

1.64

 steak

1.63

 burger

1.61

 fries

1.59

 salad

1.57

 roasted

1.56

Activations Density 0.996%

mentions or descriptions of food items and dining experiences

No Comments

No Known Activations