INDEX

Explanations

words related to unpopular opinions or positions

oai_token-act-pair · gpt-3.5-turbo

terms related to unpopularity or negative perceptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

chn

-0.85

urers

-0.84

hens

-0.77

erning

-0.74

uncture

-0.72

EStreamFrame

-0.72

ramid

-0.72

fleet

-0.70

cell

-0.69

urer

-0.69

POSITIVE LOGITS

 unpopular

1.17

ity

1.16

 incumbent

0.89

ities

0.87

liest

0.78

 disadvant

0.77

 burdens

0.77

 majorities

0.73

 incumb

0.73

 enemy

0.72

Activations Density 0.011%

words related to unpopular opinions or positions

terms related to unpopularity or negative perceptions

No Comments

No Known Activations