INDEX

Explanations

words related to personal preferences or choices

oai_token-act-pair · gpt-3.5-turbo

instances of the word "favorites" along with related discussions about preferences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ynthesis

-0.75

yre

-0.70

rogram

-0.65

ynt

-0.63

ploy

-0.63

Act

-0.62

yan

-0.61

orous

-0.60

EMENT

-0.59

aum

-0.59

POSITIVE LOGITS

 favorites

3.90

 favourites

3.40

 contenders

1.61

 Favor

1.58

 classics

1.55

 staples

1.42

 selections

1.34

 originals

1.33

 finalists

1.31

favorite

1.31

Activations Density 0.027%

words related to personal preferences or choices

instances of the word "favorites" along with related discussions about preferences

No Comments

No Known Activations