INDEX

Explanations

phrases related to showing preference or support towards something or someone

oai_token-act-pair · gpt-3.5-turbo

expressions related to favoritism or preference

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

att

-0.84

sis

-0.83

sen

-0.78

pt

-0.77

mberg

-0.77

Ä±

-0.77

ident

-0.77

thur

-0.76

ft

-0.76

hid

-0.76

POSITIVE LOGITS

 favored

1.37

 favoured

1.11

 favoring

1.09

itism

1.02

 favors

0.97

 favorites

0.93

nesday

0.88

fav

0.85

 favorable

0.81

 favorite

0.80

Activations Density 0.005%

phrases related to showing preference or support towards something or someone

expressions related to favoritism or preference

No Comments

No Known Activations