INDEX

Explanations

data-related and statistic terms

oai_token-act-pair · gpt-3.5-turbo

numerical data related to statistics and factual occurrences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

odox

-0.99

sonian

-0.72

âĵĺ

-0.66

andr

-0.65

hod

-0.64

Leader

-0.63

etta

-0.62

leader

-0.62

aja

-0.60

bart

-0.60

POSITIVE LOGITS

 worldwide

1.29

 globally

1.14

 annually

1.13

 nationwide

1.04

 yearly

0.86

 total

0.86

 domestically

0.80

 registered

0.79

 apiece

0.78

per

0.78

Activations Density 0.583%

data-related and statistic terms

numerical data related to statistics and factual occurrences

No Comments

No Known Activations