INDEX

Explanations

specific ordinal numbers ending in '-th'

oai_token-act-pair · gpt-3.5-turbo

terms related to first-round selections in various contexts

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

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

 brim

-0.60

 mars

-0.60

 formula

-0.59

Ń·

-0.59

 Robertson

-0.59

 Tant

-0.58

 shroud

-0.58

 Noir

-0.56

 Rath

-0.55

PD

-0.54

POSITIVE LOGITS

degree

1.29

generation

1.28

ever

1.28

round

1.20

century

1.17

tier

1.14

class

1.10

rate

1.09

timer

1.09

party

1.07

Activations Density 0.020%

specific ordinal numbers ending in '-th'

terms related to first-round selections in various contexts

No Comments

No Known Activations