INDEX

Explanations

phrases or words indicating the "best" or "top" choice or option

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate rankings or comparisons

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

factor

-0.68

krit

-0.64

shell

-0.64

hyde

-0.62

ADD

-0.61

ãĥ£

-0.59

taking

-0.59

kamp

-0.59

nsic

-0.59

¿

-0.59

POSITIVE LOGITS

 luck

1.05

 Worst

0.89

nesota

0.78

owitz

0.76

 breed

0.74

 intentions

0.73

 Nanto

0.70

 Luck

0.69

luck

0.69

 Practices

0.68

Activations Density 0.074%

phrases or words indicating the "best" or "top" choice or option

phrases that indicate rankings or comparisons

No Comments

No Known Activations