INDEX

Explanations

periods followed by descriptions or conclusions

oai_token-act-pair · gpt-3.5-turbo

sentences that indicate conclusions or statements of fact

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

hoe

-0.72

 inactive

-0.63

ilated

-0.61

ucer

-0.61

liest

-0.60

isine

-0.60

reet

-0.59

eteen

-0.59

 intermediate

-0.58

userc

-0.58

POSITIVE LOGITS

 Flavoring

1.07

 Whereas

1.06

 Secondly

1.05

 Thankfully

0.99

 Fortunately

0.95

 Instead

0.95

 Certainly

0.95

 Moreover

0.94

 Hence

0.94

 Luckily

0.94

Activations Density 0.681%

periods followed by descriptions or conclusions

sentences that indicate conclusions or statements of fact

No Comments

No Known Activations