INDEX

Explanations

proper nouns related to awards, titles, and achievements

oai_token-act-pair · gpt-3.5-turbo

terms related to achievements, awards, and competitions

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

ç·

-0.79

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.71

çī

-0.71

cffffcc

-0.66

cgi

-0.65

00200000

-0.63

thumbnails

-0.61

idth

-0.61

inventoryQuantity

-0.60

sonian

-0.60

POSITIVE LOGITS

hips

1.10

 apiece

1.00

 totaling

0.99

mith

0.98

 consecut

0.94

 including

0.88

 vying

0.86

 simultaneously

0.84

 plus

0.83

 imaginable

0.82

Activations Density 0.471%

proper nouns related to awards, titles, and achievements

terms related to achievements, awards, and competitions

No Comments

No Known Activations