INDEX

Explanations

positive adjectives and attributes

oai_token-act-pair · gpt-3.5-turbo

descriptors of quality and popularity

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

helm

-0.86

vier

-0.82

orthy

-0.80

reasonable

-0.79

SPONSORED

-0.79

similar

-0.77

utics

-0.76

biased

-0.76

illance

-0.75

serious

-0.74

POSITIVE LOGITS

USS

0.93

 confines

0.88

 Crimson

0.84

liest

0.81

 Millennium

0.79

 Dakota

0.79

 Falk

0.77

 Trinity

0.77

 Notting

0.77

 Golden

0.76

Activations Density 0.264%

positive adjectives and attributes

descriptors of quality and popularity

No Comments

No Known Activations