INDEX

Explanations

mentions of different variants

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "variant" and its related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ħĭ

-0.80

yer

-0.76

Ķ

-0.73

usalem

-0.72

olulu

-0.71

yers

-0.71

itness

-0.70

cept

-0.68

ashington

-0.67

mberg

-0.66

POSITIVE LOGITS

 variants

1.30

 variant

1.24

 allele

0.94

 variations

0.87

 versions

0.85

 surn

0.83

vari

0.78

 costumes

0.78

 variation

0.77

ulously

0.75

Activations Density 0.007%

mentions of different variants

occurrences of the word "variant" and its related forms

No Comments

No Known Activations