INDEX

Explanations

terms derived from other terms or origins

oai_token-act-pair · gpt-3.5-turbo

terms related to the source or origin of something

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

ebus

-0.83

phis

-0.70

icon

-0.65

roller

-0.62

 applauded

-0.62

secut

-0.59

 shuffle

-0.58

emo

-0.58

erness

-0.57

shr

-0.57

POSITIVE LOGITS

from

0.93

 from

0.89

From

0.86

 FROM

0.84

 From

0.82

actionDate

0.80

BuyableInstoreAndOnline

0.75

 derive

0.75

Orig

0.74

minist

0.74

Activations Density 0.082%

terms derived from other terms or origins

terms related to the source or origin of something

No Comments

No Known Activations