INDEX

Explanations

locations or proper nouns related to specific places

oai_token-act-pair · gpt-3.5-turbo

references to specific geographic locations and names associated with them

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

=-=-=-=-

-0.77

=-=-=-=-=-=-=-=-

-0.70

 markup

-0.66

ffen

-0.63

 flyers

-0.63

 impunity

-0.61

 upward

-0.60

orer

-0.60

eger

-0.60

 calcul

-0.59

POSITIVE LOGITS

 Cumber

1.35

batch

1.22

maid

0.87

forth

0.86

natureconservancy

0.85

ettes

0.83

ville

0.83

hunt

0.83

umber

0.82

itial

0.80

Activations Density 0.039%

locations or proper nouns related to specific places

references to specific geographic locations and names associated with them

No Comments

No Known Activations