INDEX

Explanations

references to individuals or entities described as respected

oai_token-act-pair · gpt-3.5-turbo

words indicating high levels of respect or esteem for individuals or groups

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

seed

-1.00

adra

-0.77

arettes

-0.76

ohyd

-0.74

plet

-0.74

opter

-0.73

eper

-0.72

plant

-0.72

plane

-0.70

door

-0.70

POSITIVE LOGITS

 respected

0.89

 renowned

0.83

 prestigious

0.81

respected

0.79

 reputable

0.78

 citiz

0.77

 journalistic

0.76

 scholarly

0.75

 venerable

0.74

 reputation

0.74

Activations Density 0.034%

references to individuals or entities described as respected

words indicating high levels of respect or esteem for individuals or groups

No Comments

No Known Activations