INDEX

Explanations

terms related to opinions or viewpoints

oai_token-act-pair · gpt-3.5-turbo

expressions related to opinions and perspectives

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

trap

-0.76

 Dull

-0.71

insert

-0.66

 filler

-0.65

eri

-0.65

gallery

-0.64

ombies

-0.64

wealth

-0.64

inse

-0.64

 unbeliev

-0.64

POSITIVE LOGITS

 stances

1.13

 regarding

1.07

odox

1.02

 stance

1.00

 concerning

0.99

 toward

0.93

 towards

0.92

 views

0.91

 opinions

0.91

 beliefs

0.91

Activations Density 0.108%

terms related to opinions or viewpoints

expressions related to opinions and perspectives

No Comments

No Known Activations