INDEX

Explanations

expressions of personal feelings or opinions

oai_token-act-pair · gpt-3.5-turbo

expressions of personal feelings and perspectives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

etts

-0.91

 Fahrenheit

-0.65

 Melania

-0.63

FEC

-0.62

 Marriott

-0.61

 Feast

-0.60

 Burgess

-0.59

adan

-0.58

enson

-0.58

 Debor

-0.58

POSITIVE LOGITS

 unfairly

1.09

 discriminated

0.92

 deserved

0.89

 betrayed

0.87

Interstitial

0.87

 kindred

0.87

 inadequ

0.84

 unfair

0.84

vind

0.83

 injustice

0.82

Activations Density 0.279%

expressions of personal feelings or opinions

expressions of personal feelings and perspectives

No Comments

No Known Activations