INDEX

Explanations

comparisons or evaluations based on perception

oai_token-act-pair · gpt-3.5-turbo

phrases that express skepticism or challenge perceptions

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

anwhile

-0.86

"]=>

-0.83

 Preferred

-0.76

etheless

-0.76

cised

-0.76

ESE

-0.70

aughtered

-0.69

ardless

-0.68

 WATCH

-0.68

rower

-0.68

POSITIVE LOGITS

 daunting

1.22

 innocuous

1.15

 trivial

1.13

 clichÃ©

1.10

 intimidating

1.07

 intuitive

1.06

 insignificant

1.06

 obvious

1.05

 simplistic

1.05

 outlandish

1.01

Activations Density 0.194%

comparisons or evaluations based on perception

phrases that express skepticism or challenge perceptions

No Comments

No Known Activations