INDEX

Explanations

phrases indicating the beginning or completion of a task or discussion

oai_token-act-pair · gpt-3.5-turbo

phrases related to being out of the ordinary or unconventional situations

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

utical

-0.69

raud

-0.66

cum

-0.65

ãĥį

-0.65

lege

-0.63

ILA

-0.63

ume

-0.62

OK

-0.61

oster

-0.61

olo

-0.61

POSITIVE LOGITS

 equation

1.02

 gate

0.98

 frying

0.90

 closet

0.89

 loop

0.84

box

0.84

 gates

0.83

 woods

0.81

 realm

0.80

 fray

0.77

Activations Density 0.063%

phrases indicating the beginning or completion of a task or discussion

phrases related to being out of the ordinary or unconventional situations

No Comments

No Known Activations